1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
14 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
15 Free Software Foundation, Inc.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.2 or
19 any later version published by the Free Software Foundation; with the
20 Invariant Sections being ``GNU General Public License'' and ``Funding
21 Free Software'', the Front-Cover texts being (a) (see below), and with
22 the Back-Cover Texts being (b) (see below). A copy of the license is
23 included in the gfdl(7) man page.
25 (a) The FSF's Front-Cover Text is:
29 (b) The FSF's Back-Cover Text is:
31 You have freedom to copy and modify this GNU Manual, like GNU
32 software. Copies published by the Free Software Foundation raise
33 funds for GNU development.
35 @c Set file name and title for the man page.
37 @settitle GNU project C and C++ compiler
39 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
40 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
41 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
42 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
43 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
44 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
45 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
47 Only the most useful options are listed here; see below for the
48 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
51 gpl(7), gfdl(7), fsf-funding(7),
52 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
53 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
54 @file{ld}, @file{binutils} and @file{gdb}.
57 For instructions on reporting bugs, see
61 See the Info entry for @command{gcc}, or
62 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
63 for contributors to GCC@.
68 @chapter GCC Command Options
69 @cindex GCC command options
70 @cindex command options
71 @cindex options, GCC command
73 @c man begin DESCRIPTION
74 When you invoke GCC, it normally does preprocessing, compilation,
75 assembly and linking. The ``overall options'' allow you to stop this
76 process at an intermediate stage. For example, the @option{-c} option
77 says not to run the linker. Then the output consists of object files
78 output by the assembler.
80 Other options are passed on to one stage of processing. Some options
81 control the preprocessor and others the compiler itself. Yet other
82 options control the assembler and linker; most of these are not
83 documented here, since you rarely need to use any of them.
85 @cindex C compilation options
86 Most of the command line options that you can use with GCC are useful
87 for C programs; when an option is only useful with another language
88 (usually C++), the explanation says so explicitly. If the description
89 for a particular option does not mention a source language, you can use
90 that option with all supported languages.
92 @cindex C++ compilation options
93 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
94 options for compiling C++ programs.
96 @cindex grouping options
97 @cindex options, grouping
98 The @command{gcc} program accepts options and file names as operands. Many
99 options have multi-letter names; therefore multiple single-letter options
100 may @emph{not} be grouped: @option{-dv} is very different from @w{@samp{-d
103 @cindex order of options
104 @cindex options, order
105 You can mix options and other arguments. For the most part, the order
106 you use doesn't matter. Order does matter when you use several
107 options of the same kind; for example, if you specify @option{-L} more
108 than once, the directories are searched in the order specified. Also,
109 the placement of the @option{-l} option is significant.
111 Many options have long names starting with @samp{-f} or with
112 @samp{-W}---for example,
113 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
114 these have both positive and negative forms; the negative form of
115 @option{-ffoo} would be @option{-fno-foo}. This manual documents
116 only one of these two forms, whichever one is not the default.
120 @xref{Option Index}, for an index to GCC's options.
123 * Option Summary:: Brief list of all options, without explanations.
124 * Overall Options:: Controlling the kind of output:
125 an executable, object files, assembler files,
126 or preprocessed source.
127 * Invoking G++:: Compiling C++ programs.
128 * C Dialect Options:: Controlling the variant of C language compiled.
129 * C++ Dialect Options:: Variations on C++.
130 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
132 * Language Independent Options:: Controlling how diagnostics should be
134 * Warning Options:: How picky should the compiler be?
135 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
136 * Optimize Options:: How much optimization?
137 * Preprocessor Options:: Controlling header files and macro definitions.
138 Also, getting dependency information for Make.
139 * Assembler Options:: Passing options to the assembler.
140 * Link Options:: Specifying libraries and so on.
141 * Directory Options:: Where to find header files and libraries.
142 Where to find the compiler executable files.
143 * Spec Files:: How to pass switches to sub-processes.
144 * Target Options:: Running a cross-compiler, or an old version of GCC.
145 * Submodel Options:: Specifying minor hardware or convention variations,
146 such as 68010 vs 68020.
147 * Code Gen Options:: Specifying conventions for function calls, data layout
149 * Environment Variables:: Env vars that affect GCC.
150 * Precompiled Headers:: Compiling a header once, and using it many times.
156 @section Option Summary
158 Here is a summary of all the options, grouped by type. Explanations are
159 in the following sections.
162 @item Overall Options
163 @xref{Overall Options,,Options Controlling the Kind of Output}.
164 @gccoptlist{-c -S -E -o @var{file} -combine -no-canonical-prefixes @gol
165 -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help @gol
167 --version -wrapper@@@var{file} -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg}}
169 @item C Language Options
170 @xref{C Dialect Options,,Options Controlling C Dialect}.
171 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
172 -aux-info @var{filename} @gol
173 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
174 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
175 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
176 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
177 -fsigned-bitfields -fsigned-char @gol
178 -funsigned-bitfields -funsigned-char}
180 @item C++ Language Options
181 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
182 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
183 -fconserve-space -ffriend-injection @gol
184 -fno-elide-constructors @gol
185 -fno-enforce-eh-specs @gol
186 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
187 -fno-implicit-templates @gol
188 -fno-implicit-inline-templates @gol
189 -fno-implement-inlines -fms-extensions @gol
190 -fno-nonansi-builtins -fno-operator-names @gol
191 -fno-optional-diags -fpermissive @gol
192 -fno-pretty-templates @gol
193 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
194 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
195 -fno-default-inline -fvisibility-inlines-hidden @gol
196 -fvisibility-ms-compat @gol
197 -Wabi -Wctor-dtor-privacy @gol
198 -Wnon-virtual-dtor -Wreorder @gol
199 -Weffc++ -Wstrict-null-sentinel @gol
200 -Wno-non-template-friend -Wold-style-cast @gol
201 -Woverloaded-virtual -Wno-pmf-conversions @gol
204 @item Objective-C and Objective-C++ Language Options
205 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
206 Objective-C and Objective-C++ Dialects}.
207 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
208 -fgnu-runtime -fnext-runtime @gol
209 -fno-nil-receivers @gol
210 -fobjc-call-cxx-cdtors @gol
211 -fobjc-direct-dispatch @gol
212 -fobjc-exceptions @gol
214 -freplace-objc-classes @gol
217 -Wassign-intercept @gol
218 -Wno-protocol -Wselector @gol
219 -Wstrict-selector-match @gol
220 -Wundeclared-selector}
222 @item Language Independent Options
223 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
224 @gccoptlist{-fmessage-length=@var{n} @gol
225 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
226 -fdiagnostics-show-option}
228 @item Warning Options
229 @xref{Warning Options,,Options to Request or Suppress Warnings}.
230 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
231 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
232 -Wno-attributes -Wno-builtin-macro-redefined @gol
233 -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
234 -Wchar-subscripts -Wclobbered -Wcomment @gol
235 -Wconversion -Wcoverage-mismatch -Wno-deprecated @gol
236 -Wno-deprecated-declarations -Wdisabled-optimization @gol
237 -Wno-div-by-zero -Wempty-body -Wenum-compare -Wno-endif-labels @gol
238 -Werror -Werror=* @gol
239 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
240 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
241 -Wformat-security -Wformat-y2k @gol
242 -Wframe-larger-than=@var{len} -Wjump-misses-init -Wignored-qualifiers @gol
243 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
244 -Winit-self -Winline @gol
245 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
246 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
247 -Wlogical-op -Wlong-long @gol
248 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
249 -Wmissing-format-attribute -Wmissing-include-dirs @gol
250 -Wmissing-noreturn -Wno-mudflap @gol
251 -Wno-multichar -Wnonnull -Wno-overflow @gol
252 -Woverlength-strings -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
253 -Wparentheses -Wpedantic-ms-format -Wno-pedantic-ms-format @gol
254 -Wpointer-arith -Wno-pointer-to-int-cast @gol
255 -Wredundant-decls @gol
256 -Wreturn-type -Wsequence-point -Wshadow @gol
257 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
258 -Wstrict-aliasing -Wstrict-aliasing=n @gol
259 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
260 -Wswitch -Wswitch-default -Wswitch-enum -Wsync-nand @gol
261 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
262 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
263 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
264 -Wunused-label -Wunused-parameter -Wno-unused-result -Wunused-value -Wunused-variable @gol
265 -Wvariadic-macros -Wvla @gol
266 -Wvolatile-register-var -Wwrite-strings}
268 @item C and Objective-C-only Warning Options
269 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
270 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
271 -Wold-style-declaration -Wold-style-definition @gol
272 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
273 -Wdeclaration-after-statement -Wpointer-sign}
275 @item Debugging Options
276 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
277 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
278 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
279 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
280 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
281 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
282 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
283 -fdump-statistics @gol
285 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
289 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
294 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
296 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
298 -fdump-tree-nrv -fdump-tree-vect @gol
299 -fdump-tree-sink @gol
300 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
301 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
302 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
303 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
304 -ftree-vectorizer-verbose=@var{n} @gol
305 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
306 -fdump-final-insns=@var{file} @gol
307 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
308 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
309 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
310 -fenable-icf-debug @gol
311 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
312 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
313 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
314 -ftest-coverage -ftime-report -fvar-tracking @gol
315 -fvar-tracking-assigments -fvar-tracking-assignments-toggle @gol
316 -g -g@var{level} -gtoggle -gcoff -gdwarf-@var{version} @gol
317 -ggdb -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
318 -gvms -gxcoff -gxcoff+ @gol
319 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
320 -fdebug-prefix-map=@var{old}=@var{new} @gol
321 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
322 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
323 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
324 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
325 -print-prog-name=@var{program} -print-search-dirs -Q @gol
326 -print-sysroot -print-sysroot-headers-suffix @gol
327 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
329 @item Optimization Options
330 @xref{Optimize Options,,Options that Control Optimization}.
332 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
333 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
334 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
335 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
336 -fcheck-data-deps -fconserve-stack -fcprop-registers -fcrossjumping @gol
337 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
338 -fdata-sections -fdce -fdce @gol
339 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
340 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffast-math @gol
341 -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
342 -fforward-propagate -ffunction-sections @gol
343 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
344 -fgcse-sm -fif-conversion -fif-conversion2 -findirect-inlining @gol
345 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
346 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-matrix-reorg -fipa-pta @gol
347 -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
348 -fipa-type-escape -fira-algorithm=@var{algorithm} @gol
349 -fira-region=@var{region} -fira-coalesce @gol
350 -fira-loop-pressure -fno-ira-share-save-slots @gol
351 -fno-ira-share-spill-slots -fira-verbose=@var{n} @gol
352 -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
353 -floop-block -floop-interchange -floop-strip-mine -fgraphite-identity @gol
354 -floop-parallelize-all -flto -flto-compression-level -flto-report -fltrans @gol
355 -fltrans-output-list -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
356 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
357 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
358 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
359 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
360 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
361 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
362 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
363 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
364 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
365 -fprofile-generate=@var{path} @gol
366 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
367 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
368 -freorder-blocks-and-partition -freorder-functions @gol
369 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
370 -frounding-math -fsched2-use-superblocks @gol
371 -fsched2-use-traces -fsched-pressure @gol
372 -fsched-spec-load -fsched-spec-load-dangerous @gol
373 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
374 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
375 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
376 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
377 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
378 -fselective-scheduling -fselective-scheduling2 @gol
379 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
380 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
381 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
382 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer @gol
383 -ftree-builtin-call-dce -ftree-ccp -ftree-ch -ftree-copy-prop @gol
384 -ftree-copyrename -ftree-dce @gol
385 -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre -ftree-loop-im @gol
386 -ftree-phiprop -ftree-loop-distribution @gol
387 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
388 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-pta -ftree-reassoc @gol
389 -ftree-sink -ftree-sra -ftree-switch-conversion @gol
390 -ftree-ter -ftree-vect-loop-version -ftree-vectorize -ftree-vrp @gol
391 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
392 -funsafe-loop-optimizations -funsafe-math-optimizations -funswitch-loops @gol
393 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
394 -fwhole-program -fwhopr -fwpa -use-linker-plugin @gol
395 --param @var{name}=@var{value}
396 -O -O0 -O1 -O2 -O3 -Os}
398 @item Preprocessor Options
399 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
400 @gccoptlist{-A@var{question}=@var{answer} @gol
401 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
402 -C -dD -dI -dM -dN @gol
403 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
404 -idirafter @var{dir} @gol
405 -include @var{file} -imacros @var{file} @gol
406 -iprefix @var{file} -iwithprefix @var{dir} @gol
407 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
408 -imultilib @var{dir} -isysroot @var{dir} @gol
409 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
410 -P -fworking-directory -remap @gol
411 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
412 -Xpreprocessor @var{option}}
414 @item Assembler Option
415 @xref{Assembler Options,,Passing Options to the Assembler}.
416 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
419 @xref{Link Options,,Options for Linking}.
420 @gccoptlist{@var{object-file-name} -l@var{library} @gol
421 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
422 -s -static -static-libgcc -static-libstdc++ -shared @gol
423 -shared-libgcc -symbolic @gol
424 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
427 @item Directory Options
428 @xref{Directory Options,,Options for Directory Search}.
429 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
430 -specs=@var{file} -I- --sysroot=@var{dir}}
433 @c I wrote this xref this way to avoid overfull hbox. -- rms
434 @xref{Target Options}.
435 @gccoptlist{-V @var{version} -b @var{machine}}
437 @item Machine Dependent Options
438 @xref{Submodel Options,,Hardware Models and Configurations}.
439 @c This list is ordered alphanumerically by subsection name.
440 @c Try and put the significant identifier (CPU or system) first,
441 @c so users have a clue at guessing where the ones they want will be.
444 @gccoptlist{-EB -EL @gol
445 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
446 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
449 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
450 -mabi=@var{name} @gol
451 -mapcs-stack-check -mno-apcs-stack-check @gol
452 -mapcs-float -mno-apcs-float @gol
453 -mapcs-reentrant -mno-apcs-reentrant @gol
454 -msched-prolog -mno-sched-prolog @gol
455 -mlittle-endian -mbig-endian -mwords-little-endian @gol
456 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
457 -mfp16-format=@var{name}
458 -mthumb-interwork -mno-thumb-interwork @gol
459 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
460 -mstructure-size-boundary=@var{n} @gol
461 -mabort-on-noreturn @gol
462 -mlong-calls -mno-long-calls @gol
463 -msingle-pic-base -mno-single-pic-base @gol
464 -mpic-register=@var{reg} @gol
465 -mnop-fun-dllimport @gol
466 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
467 -mpoke-function-name @gol
469 -mtpcs-frame -mtpcs-leaf-frame @gol
470 -mcaller-super-interworking -mcallee-super-interworking @gol
472 -mword-relocations @gol
473 -mfix-cortex-m3-ldrd}
476 @gccoptlist{-mmcu=@var{mcu} -mno-interrupts @gol
477 -mcall-prologues -mtiny-stack -mint8}
479 @emph{Blackfin Options}
480 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
481 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
482 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
483 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
484 -mno-id-shared-library -mshared-library-id=@var{n} @gol
485 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
486 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
487 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
491 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
492 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
493 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
494 -mstack-align -mdata-align -mconst-align @gol
495 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
496 -melf -maout -melinux -mlinux -sim -sim2 @gol
497 -mmul-bug-workaround -mno-mul-bug-workaround}
500 @gccoptlist{-mmac -mpush-args}
502 @emph{Darwin Options}
503 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
504 -arch_only -bind_at_load -bundle -bundle_loader @gol
505 -client_name -compatibility_version -current_version @gol
507 -dependency-file -dylib_file -dylinker_install_name @gol
508 -dynamic -dynamiclib -exported_symbols_list @gol
509 -filelist -flat_namespace -force_cpusubtype_ALL @gol
510 -force_flat_namespace -headerpad_max_install_names @gol
512 -image_base -init -install_name -keep_private_externs @gol
513 -multi_module -multiply_defined -multiply_defined_unused @gol
514 -noall_load -no_dead_strip_inits_and_terms @gol
515 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
516 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
517 -private_bundle -read_only_relocs -sectalign @gol
518 -sectobjectsymbols -whyload -seg1addr @gol
519 -sectcreate -sectobjectsymbols -sectorder @gol
520 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
521 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
522 -segprot -segs_read_only_addr -segs_read_write_addr @gol
523 -single_module -static -sub_library -sub_umbrella @gol
524 -twolevel_namespace -umbrella -undefined @gol
525 -unexported_symbols_list -weak_reference_mismatches @gol
526 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
527 -mkernel -mone-byte-bool}
529 @emph{DEC Alpha Options}
530 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
531 -mieee -mieee-with-inexact -mieee-conformant @gol
532 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
533 -mtrap-precision=@var{mode} -mbuild-constants @gol
534 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
535 -mbwx -mmax -mfix -mcix @gol
536 -mfloat-vax -mfloat-ieee @gol
537 -mexplicit-relocs -msmall-data -mlarge-data @gol
538 -msmall-text -mlarge-text @gol
539 -mmemory-latency=@var{time}}
541 @emph{DEC Alpha/VMS Options}
542 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
545 @gccoptlist{-msmall-model -mno-lsim}
548 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
549 -mhard-float -msoft-float @gol
550 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
551 -mdouble -mno-double @gol
552 -mmedia -mno-media -mmuladd -mno-muladd @gol
553 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
554 -mlinked-fp -mlong-calls -malign-labels @gol
555 -mlibrary-pic -macc-4 -macc-8 @gol
556 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
557 -moptimize-membar -mno-optimize-membar @gol
558 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
559 -mvliw-branch -mno-vliw-branch @gol
560 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
561 -mno-nested-cond-exec -mtomcat-stats @gol
565 @emph{GNU/Linux Options}
566 @gccoptlist{-muclibc}
568 @emph{H8/300 Options}
569 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
572 @gccoptlist{-march=@var{architecture-type} @gol
573 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
574 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
575 -mfixed-range=@var{register-range} @gol
576 -mjump-in-delay -mlinker-opt -mlong-calls @gol
577 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
578 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
579 -mno-jump-in-delay -mno-long-load-store @gol
580 -mno-portable-runtime -mno-soft-float @gol
581 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
582 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
583 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
584 -munix=@var{unix-std} -nolibdld -static -threads}
586 @emph{i386 and x86-64 Options}
587 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
588 -mfpmath=@var{unit} @gol
589 -masm=@var{dialect} -mno-fancy-math-387 @gol
590 -mno-fp-ret-in-387 -msoft-float @gol
591 -mno-wide-multiply -mrtd -malign-double @gol
592 -mpreferred-stack-boundary=@var{num}
593 -mincoming-stack-boundary=@var{num}
594 -mcld -mcx16 -msahf -mmovbe -mcrc32 -mrecip @gol
595 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
597 -msse4a -m3dnow -mpopcnt -mabm -mfma4 @gol
598 -mthreads -mno-align-stringops -minline-all-stringops @gol
599 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
600 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
601 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
602 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
603 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
604 -mcmodel=@var{code-model} -mabi=@var{name} @gol
605 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
609 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
610 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
611 -mconstant-gp -mauto-pic -mfused-madd @gol
612 -minline-float-divide-min-latency @gol
613 -minline-float-divide-max-throughput @gol
614 -mno-inline-float-divide @gol
615 -minline-int-divide-min-latency @gol
616 -minline-int-divide-max-throughput @gol
617 -mno-inline-int-divide @gol
618 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
619 -mno-inline-sqrt @gol
620 -mdwarf2-asm -mearly-stop-bits @gol
621 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
622 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
623 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
624 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
625 -msched-spec-ldc -msched-spec-control-ldc @gol
626 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
627 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
628 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
629 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
631 @emph{IA-64/VMS Options}
632 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
634 @emph{M32R/D Options}
635 @gccoptlist{-m32r2 -m32rx -m32r @gol
637 -malign-loops -mno-align-loops @gol
638 -missue-rate=@var{number} @gol
639 -mbranch-cost=@var{number} @gol
640 -mmodel=@var{code-size-model-type} @gol
641 -msdata=@var{sdata-type} @gol
642 -mno-flush-func -mflush-func=@var{name} @gol
643 -mno-flush-trap -mflush-trap=@var{number} @gol
647 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
649 @emph{M680x0 Options}
650 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
651 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
652 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
653 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
654 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
655 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
656 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
657 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
660 @emph{M68hc1x Options}
661 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
662 -mauto-incdec -minmax -mlong-calls -mshort @gol
663 -msoft-reg-count=@var{count}}
666 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
667 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
668 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
669 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
670 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
673 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
674 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
675 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
676 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
680 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
681 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 @gol
682 -mips64 -mips64r2 @gol
683 -mips16 -mno-mips16 -mflip-mips16 @gol
684 -minterlink-mips16 -mno-interlink-mips16 @gol
685 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
686 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
687 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
688 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
689 -mfpu=@var{fpu-type} @gol
690 -msmartmips -mno-smartmips @gol
691 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
692 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
693 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
694 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
695 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
696 -membedded-data -mno-embedded-data @gol
697 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
698 -mcode-readable=@var{setting} @gol
699 -msplit-addresses -mno-split-addresses @gol
700 -mexplicit-relocs -mno-explicit-relocs @gol
701 -mcheck-zero-division -mno-check-zero-division @gol
702 -mdivide-traps -mdivide-breaks @gol
703 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
704 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
705 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
706 -mfix-r10000 -mno-fix-r10000 -mfix-vr4120 -mno-fix-vr4120 @gol
707 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
708 -mflush-func=@var{func} -mno-flush-func @gol
709 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
710 -mfp-exceptions -mno-fp-exceptions @gol
711 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
712 -mrelax-pic-calls -mno-relax-pic-calls}
715 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
716 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
717 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
718 -mno-base-addresses -msingle-exit -mno-single-exit}
720 @emph{MN10300 Options}
721 @gccoptlist{-mmult-bug -mno-mult-bug @gol
722 -mam33 -mno-am33 @gol
723 -mam33-2 -mno-am33-2 @gol
724 -mreturn-pointer-on-d0 @gol
727 @emph{PDP-11 Options}
728 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
729 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
730 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
731 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
732 -mbranch-expensive -mbranch-cheap @gol
733 -msplit -mno-split -munix-asm -mdec-asm}
735 @emph{picoChip Options}
736 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N}
737 -msymbol-as-address -mno-inefficient-warnings}
739 @emph{PowerPC Options}
740 See RS/6000 and PowerPC Options.
742 @emph{RS/6000 and PowerPC Options}
743 @gccoptlist{-mcpu=@var{cpu-type} @gol
744 -mtune=@var{cpu-type} @gol
745 -mpower -mno-power -mpower2 -mno-power2 @gol
746 -mpowerpc -mpowerpc64 -mno-powerpc @gol
747 -maltivec -mno-altivec @gol
748 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
749 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
750 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
751 -mfprnd -mno-fprnd @gol
752 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
753 -mnew-mnemonics -mold-mnemonics @gol
754 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
755 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
756 -malign-power -malign-natural @gol
757 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
758 -msingle-float -mdouble-float -msimple-fpu @gol
759 -mstring -mno-string -mupdate -mno-update @gol
760 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
761 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
762 -mstrict-align -mno-strict-align -mrelocatable @gol
763 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
764 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
765 -mdynamic-no-pic -maltivec -mswdiv @gol
766 -mprioritize-restricted-insns=@var{priority} @gol
767 -msched-costly-dep=@var{dependence_type} @gol
768 -minsert-sched-nops=@var{scheme} @gol
769 -mcall-sysv -mcall-netbsd @gol
770 -maix-struct-return -msvr4-struct-return @gol
771 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
772 -misel -mno-isel @gol
773 -misel=yes -misel=no @gol
775 -mspe=yes -mspe=no @gol
777 -mgen-cell-microcode -mwarn-cell-microcode @gol
778 -mvrsave -mno-vrsave @gol
779 -mmulhw -mno-mulhw @gol
780 -mdlmzb -mno-dlmzb @gol
781 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
782 -mprototype -mno-prototype @gol
783 -msim -mmvme -mads -myellowknife -memb -msdata @gol
784 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
787 @gccoptlist{-m64bit-doubles -m32bit-doubles -mieee -mno-ieee@gol
788 -mbig-endian-data -mlittle-endian-data @gol
791 -mas100-syntax -mno-as100-syntax@gol
793 -mmax-constant-size=@gol
796 @emph{S/390 and zSeries Options}
797 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
798 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
799 -mlong-double-64 -mlong-double-128 @gol
800 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
801 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
802 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
803 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
804 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
807 @gccoptlist{-meb -mel @gol
811 -mscore5 -mscore5u -mscore7 -mscore7d}
814 @gccoptlist{-m1 -m2 -m2e @gol
815 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
817 -m4-nofpu -m4-single-only -m4-single -m4 @gol
818 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
819 -m5-64media -m5-64media-nofpu @gol
820 -m5-32media -m5-32media-nofpu @gol
821 -m5-compact -m5-compact-nofpu @gol
822 -mb -ml -mdalign -mrelax @gol
823 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
824 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
825 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
826 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
827 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
831 @gccoptlist{-mcpu=@var{cpu-type} @gol
832 -mtune=@var{cpu-type} @gol
833 -mcmodel=@var{code-model} @gol
834 -m32 -m64 -mapp-regs -mno-app-regs @gol
835 -mfaster-structs -mno-faster-structs @gol
836 -mfpu -mno-fpu -mhard-float -msoft-float @gol
837 -mhard-quad-float -msoft-quad-float @gol
838 -mimpure-text -mno-impure-text -mlittle-endian @gol
839 -mstack-bias -mno-stack-bias @gol
840 -munaligned-doubles -mno-unaligned-doubles @gol
841 -mv8plus -mno-v8plus -mvis -mno-vis
842 -threads -pthreads -pthread}
845 @gccoptlist{-mwarn-reloc -merror-reloc @gol
846 -msafe-dma -munsafe-dma @gol
848 -msmall-mem -mlarge-mem -mstdmain @gol
849 -mfixed-range=@var{register-range}}
851 @emph{System V Options}
852 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
855 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
856 -mprolog-function -mno-prolog-function -mspace @gol
857 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
858 -mapp-regs -mno-app-regs @gol
859 -mdisable-callt -mno-disable-callt @gol
865 @gccoptlist{-mg -mgnu -munix}
867 @emph{VxWorks Options}
868 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
869 -Xbind-lazy -Xbind-now}
871 @emph{x86-64 Options}
872 See i386 and x86-64 Options.
874 @emph{i386 and x86-64 Windows Options}
875 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll
876 -mnop-fun-dllimport -mthread -municode -mwin32 -mwindows}
878 @emph{Xstormy16 Options}
881 @emph{Xtensa Options}
882 @gccoptlist{-mconst16 -mno-const16 @gol
883 -mfused-madd -mno-fused-madd @gol
884 -mserialize-volatile -mno-serialize-volatile @gol
885 -mtext-section-literals -mno-text-section-literals @gol
886 -mtarget-align -mno-target-align @gol
887 -mlongcalls -mno-longcalls}
889 @emph{zSeries Options}
890 See S/390 and zSeries Options.
892 @item Code Generation Options
893 @xref{Code Gen Options,,Options for Code Generation Conventions}.
894 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
895 -ffixed-@var{reg} -fexceptions @gol
896 -fnon-call-exceptions -funwind-tables @gol
897 -fasynchronous-unwind-tables @gol
898 -finhibit-size-directive -finstrument-functions @gol
899 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
900 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
901 -fno-common -fno-ident @gol
902 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
903 -fno-jump-tables @gol
904 -frecord-gcc-switches @gol
905 -freg-struct-return -fshort-enums @gol
906 -fshort-double -fshort-wchar @gol
907 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
908 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
909 -fno-stack-limit -fargument-alias -fargument-noalias @gol
910 -fargument-noalias-global -fargument-noalias-anything @gol
911 -fleading-underscore -ftls-model=@var{model} @gol
912 -ftrapv -fwrapv -fbounds-check @gol
917 * Overall Options:: Controlling the kind of output:
918 an executable, object files, assembler files,
919 or preprocessed source.
920 * C Dialect Options:: Controlling the variant of C language compiled.
921 * C++ Dialect Options:: Variations on C++.
922 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
924 * Language Independent Options:: Controlling how diagnostics should be
926 * Warning Options:: How picky should the compiler be?
927 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
928 * Optimize Options:: How much optimization?
929 * Preprocessor Options:: Controlling header files and macro definitions.
930 Also, getting dependency information for Make.
931 * Assembler Options:: Passing options to the assembler.
932 * Link Options:: Specifying libraries and so on.
933 * Directory Options:: Where to find header files and libraries.
934 Where to find the compiler executable files.
935 * Spec Files:: How to pass switches to sub-processes.
936 * Target Options:: Running a cross-compiler, or an old version of GCC.
939 @node Overall Options
940 @section Options Controlling the Kind of Output
942 Compilation can involve up to four stages: preprocessing, compilation
943 proper, assembly and linking, always in that order. GCC is capable of
944 preprocessing and compiling several files either into several
945 assembler input files, or into one assembler input file; then each
946 assembler input file produces an object file, and linking combines all
947 the object files (those newly compiled, and those specified as input)
948 into an executable file.
950 @cindex file name suffix
951 For any given input file, the file name suffix determines what kind of
956 C source code which must be preprocessed.
959 C source code which should not be preprocessed.
962 C++ source code which should not be preprocessed.
965 Objective-C source code. Note that you must link with the @file{libobjc}
966 library to make an Objective-C program work.
969 Objective-C source code which should not be preprocessed.
973 Objective-C++ source code. Note that you must link with the @file{libobjc}
974 library to make an Objective-C++ program work. Note that @samp{.M} refers
975 to a literal capital M@.
978 Objective-C++ source code which should not be preprocessed.
981 C, C++, Objective-C or Objective-C++ header file to be turned into a
986 @itemx @var{file}.cxx
987 @itemx @var{file}.cpp
988 @itemx @var{file}.CPP
989 @itemx @var{file}.c++
991 C++ source code which must be preprocessed. Note that in @samp{.cxx},
992 the last two letters must both be literally @samp{x}. Likewise,
993 @samp{.C} refers to a literal capital C@.
997 Objective-C++ source code which must be preprocessed.
1000 Objective-C++ source code which should not be preprocessed.
1004 @itemx @var{file}.hp
1005 @itemx @var{file}.hxx
1006 @itemx @var{file}.hpp
1007 @itemx @var{file}.HPP
1008 @itemx @var{file}.h++
1009 @itemx @var{file}.tcc
1010 C++ header file to be turned into a precompiled header.
1013 @itemx @var{file}.for
1014 @itemx @var{file}.ftn
1015 Fixed form Fortran source code which should not be preprocessed.
1018 @itemx @var{file}.FOR
1019 @itemx @var{file}.fpp
1020 @itemx @var{file}.FPP
1021 @itemx @var{file}.FTN
1022 Fixed form Fortran source code which must be preprocessed (with the traditional
1025 @item @var{file}.f90
1026 @itemx @var{file}.f95
1027 @itemx @var{file}.f03
1028 @itemx @var{file}.f08
1029 Free form Fortran source code which should not be preprocessed.
1031 @item @var{file}.F90
1032 @itemx @var{file}.F95
1033 @itemx @var{file}.F03
1034 @itemx @var{file}.F08
1035 Free form Fortran source code which must be preprocessed (with the
1036 traditional preprocessor).
1038 @c FIXME: Descriptions of Java file types.
1044 @item @var{file}.ads
1045 Ada source code file which contains a library unit declaration (a
1046 declaration of a package, subprogram, or generic, or a generic
1047 instantiation), or a library unit renaming declaration (a package,
1048 generic, or subprogram renaming declaration). Such files are also
1051 @item @var{file}.adb
1052 Ada source code file containing a library unit body (a subprogram or
1053 package body). Such files are also called @dfn{bodies}.
1055 @c GCC also knows about some suffixes for languages not yet included:
1066 @itemx @var{file}.sx
1067 Assembler code which must be preprocessed.
1070 An object file to be fed straight into linking.
1071 Any file name with no recognized suffix is treated this way.
1075 You can specify the input language explicitly with the @option{-x} option:
1078 @item -x @var{language}
1079 Specify explicitly the @var{language} for the following input files
1080 (rather than letting the compiler choose a default based on the file
1081 name suffix). This option applies to all following input files until
1082 the next @option{-x} option. Possible values for @var{language} are:
1084 c c-header c-cpp-output
1085 c++ c++-header c++-cpp-output
1086 objective-c objective-c-header objective-c-cpp-output
1087 objective-c++ objective-c++-header objective-c++-cpp-output
1088 assembler assembler-with-cpp
1090 f77 f77-cpp-input f95 f95-cpp-input
1095 Turn off any specification of a language, so that subsequent files are
1096 handled according to their file name suffixes (as they are if @option{-x}
1097 has not been used at all).
1099 @item -pass-exit-codes
1100 @opindex pass-exit-codes
1101 Normally the @command{gcc} program will exit with the code of 1 if any
1102 phase of the compiler returns a non-success return code. If you specify
1103 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1104 numerically highest error produced by any phase that returned an error
1105 indication. The C, C++, and Fortran frontends return 4, if an internal
1106 compiler error is encountered.
1109 If you only want some of the stages of compilation, you can use
1110 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1111 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1112 @command{gcc} is to stop. Note that some combinations (for example,
1113 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1118 Compile or assemble the source files, but do not link. The linking
1119 stage simply is not done. The ultimate output is in the form of an
1120 object file for each source file.
1122 By default, the object file name for a source file is made by replacing
1123 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1125 Unrecognized input files, not requiring compilation or assembly, are
1130 Stop after the stage of compilation proper; do not assemble. The output
1131 is in the form of an assembler code file for each non-assembler input
1134 By default, the assembler file name for a source file is made by
1135 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1137 Input files that don't require compilation are ignored.
1141 Stop after the preprocessing stage; do not run the compiler proper. The
1142 output is in the form of preprocessed source code, which is sent to the
1145 Input files which don't require preprocessing are ignored.
1147 @cindex output file option
1150 Place output in file @var{file}. This applies regardless to whatever
1151 sort of output is being produced, whether it be an executable file,
1152 an object file, an assembler file or preprocessed C code.
1154 If @option{-o} is not specified, the default is to put an executable
1155 file in @file{a.out}, the object file for
1156 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1157 assembler file in @file{@var{source}.s}, a precompiled header file in
1158 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1163 Print (on standard error output) the commands executed to run the stages
1164 of compilation. Also print the version number of the compiler driver
1165 program and of the preprocessor and the compiler proper.
1169 Like @option{-v} except the commands are not executed and all command
1170 arguments are quoted. This is useful for shell scripts to capture the
1171 driver-generated command lines.
1175 Use pipes rather than temporary files for communication between the
1176 various stages of compilation. This fails to work on some systems where
1177 the assembler is unable to read from a pipe; but the GNU assembler has
1182 If you are compiling multiple source files, this option tells the driver
1183 to pass all the source files to the compiler at once (for those
1184 languages for which the compiler can handle this). This will allow
1185 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1186 language for which this is supported is C@. If you pass source files for
1187 multiple languages to the driver, using this option, the driver will invoke
1188 the compiler(s) that support IMA once each, passing each compiler all the
1189 source files appropriate for it. For those languages that do not support
1190 IMA this option will be ignored, and the compiler will be invoked once for
1191 each source file in that language. If you use this option in conjunction
1192 with @option{-save-temps}, the compiler will generate multiple
1194 (one for each source file), but only one (combined) @file{.o} or
1199 Print (on the standard output) a description of the command line options
1200 understood by @command{gcc}. If the @option{-v} option is also specified
1201 then @option{--help} will also be passed on to the various processes
1202 invoked by @command{gcc}, so that they can display the command line options
1203 they accept. If the @option{-Wextra} option has also been specified
1204 (prior to the @option{--help} option), then command line options which
1205 have no documentation associated with them will also be displayed.
1208 @opindex target-help
1209 Print (on the standard output) a description of target-specific command
1210 line options for each tool. For some targets extra target-specific
1211 information may also be printed.
1213 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1214 Print (on the standard output) a description of the command line
1215 options understood by the compiler that fit into all specified classes
1216 and qualifiers. These are the supported classes:
1219 @item @samp{optimizers}
1220 This will display all of the optimization options supported by the
1223 @item @samp{warnings}
1224 This will display all of the options controlling warning messages
1225 produced by the compiler.
1228 This will display target-specific options. Unlike the
1229 @option{--target-help} option however, target-specific options of the
1230 linker and assembler will not be displayed. This is because those
1231 tools do not currently support the extended @option{--help=} syntax.
1234 This will display the values recognized by the @option{--param}
1237 @item @var{language}
1238 This will display the options supported for @var{language}, where
1239 @var{language} is the name of one of the languages supported in this
1243 This will display the options that are common to all languages.
1246 These are the supported qualifiers:
1249 @item @samp{undocumented}
1250 Display only those options which are undocumented.
1253 Display options which take an argument that appears after an equal
1254 sign in the same continuous piece of text, such as:
1255 @samp{--help=target}.
1257 @item @samp{separate}
1258 Display options which take an argument that appears as a separate word
1259 following the original option, such as: @samp{-o output-file}.
1262 Thus for example to display all the undocumented target-specific
1263 switches supported by the compiler the following can be used:
1266 --help=target,undocumented
1269 The sense of a qualifier can be inverted by prefixing it with the
1270 @samp{^} character, so for example to display all binary warning
1271 options (i.e., ones that are either on or off and that do not take an
1272 argument), which have a description the following can be used:
1275 --help=warnings,^joined,^undocumented
1278 The argument to @option{--help=} should not consist solely of inverted
1281 Combining several classes is possible, although this usually
1282 restricts the output by so much that there is nothing to display. One
1283 case where it does work however is when one of the classes is
1284 @var{target}. So for example to display all the target-specific
1285 optimization options the following can be used:
1288 --help=target,optimizers
1291 The @option{--help=} option can be repeated on the command line. Each
1292 successive use will display its requested class of options, skipping
1293 those that have already been displayed.
1295 If the @option{-Q} option appears on the command line before the
1296 @option{--help=} option, then the descriptive text displayed by
1297 @option{--help=} is changed. Instead of describing the displayed
1298 options, an indication is given as to whether the option is enabled,
1299 disabled or set to a specific value (assuming that the compiler
1300 knows this at the point where the @option{--help=} option is used).
1302 Here is a truncated example from the ARM port of @command{gcc}:
1305 % gcc -Q -mabi=2 --help=target -c
1306 The following options are target specific:
1308 -mabort-on-noreturn [disabled]
1312 The output is sensitive to the effects of previous command line
1313 options, so for example it is possible to find out which optimizations
1314 are enabled at @option{-O2} by using:
1317 -Q -O2 --help=optimizers
1320 Alternatively you can discover which binary optimizations are enabled
1321 by @option{-O3} by using:
1324 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1325 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1326 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1329 @item -no-canonical-prefixes
1330 @opindex no-canonical-prefixes
1331 Do not expand any symbolic links, resolve references to @samp{/../}
1332 or @samp{/./}, or make the path absolute when generating a relative
1337 Display the version number and copyrights of the invoked GCC@.
1341 Invoke all subcommands under a wrapper program. It takes a single
1342 comma separated list as an argument, which will be used to invoke
1346 gcc -c t.c -wrapper gdb,--args
1349 This will invoke all subprograms of gcc under "gdb --args",
1350 thus cc1 invocation will be "gdb --args cc1 ...".
1352 @item -fplugin=@var{name}.so
1353 Load the plugin code in file @var{name}.so, assumed to be a
1354 shared object to be dlopen'd by the compiler. The base name of
1355 the shared object file is used to identify the plugin for the
1356 purposes of argument parsing (See
1357 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1358 Each plugin should define the callback functions specified in the
1361 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1362 Define an argument called @var{key} with a value of @var{value}
1363 for the plugin called @var{name}.
1365 @include @value{srcdir}/../libiberty/at-file.texi
1369 @section Compiling C++ Programs
1371 @cindex suffixes for C++ source
1372 @cindex C++ source file suffixes
1373 C++ source files conventionally use one of the suffixes @samp{.C},
1374 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1375 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1376 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1377 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1378 files with these names and compiles them as C++ programs even if you
1379 call the compiler the same way as for compiling C programs (usually
1380 with the name @command{gcc}).
1384 However, the use of @command{gcc} does not add the C++ library.
1385 @command{g++} is a program that calls GCC and treats @samp{.c},
1386 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1387 files unless @option{-x} is used, and automatically specifies linking
1388 against the C++ library. This program is also useful when
1389 precompiling a C header file with a @samp{.h} extension for use in C++
1390 compilations. On many systems, @command{g++} is also installed with
1391 the name @command{c++}.
1393 @cindex invoking @command{g++}
1394 When you compile C++ programs, you may specify many of the same
1395 command-line options that you use for compiling programs in any
1396 language; or command-line options meaningful for C and related
1397 languages; or options that are meaningful only for C++ programs.
1398 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1399 explanations of options for languages related to C@.
1400 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1401 explanations of options that are meaningful only for C++ programs.
1403 @node C Dialect Options
1404 @section Options Controlling C Dialect
1405 @cindex dialect options
1406 @cindex language dialect options
1407 @cindex options, dialect
1409 The following options control the dialect of C (or languages derived
1410 from C, such as C++, Objective-C and Objective-C++) that the compiler
1414 @cindex ANSI support
1418 In C mode, this is equivalent to @samp{-std=c89}. In C++ mode, it is
1419 equivalent to @samp{-std=c++98}.
1421 This turns off certain features of GCC that are incompatible with ISO
1422 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1423 such as the @code{asm} and @code{typeof} keywords, and
1424 predefined macros such as @code{unix} and @code{vax} that identify the
1425 type of system you are using. It also enables the undesirable and
1426 rarely used ISO trigraph feature. For the C compiler,
1427 it disables recognition of C++ style @samp{//} comments as well as
1428 the @code{inline} keyword.
1430 The alternate keywords @code{__asm__}, @code{__extension__},
1431 @code{__inline__} and @code{__typeof__} continue to work despite
1432 @option{-ansi}. You would not want to use them in an ISO C program, of
1433 course, but it is useful to put them in header files that might be included
1434 in compilations done with @option{-ansi}. Alternate predefined macros
1435 such as @code{__unix__} and @code{__vax__} are also available, with or
1436 without @option{-ansi}.
1438 The @option{-ansi} option does not cause non-ISO programs to be
1439 rejected gratuitously. For that, @option{-pedantic} is required in
1440 addition to @option{-ansi}. @xref{Warning Options}.
1442 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1443 option is used. Some header files may notice this macro and refrain
1444 from declaring certain functions or defining certain macros that the
1445 ISO standard doesn't call for; this is to avoid interfering with any
1446 programs that might use these names for other things.
1448 Functions that would normally be built in but do not have semantics
1449 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1450 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1451 built-in functions provided by GCC}, for details of the functions
1456 Determine the language standard. @xref{Standards,,Language Standards
1457 Supported by GCC}, for details of these standard versions. This option
1458 is currently only supported when compiling C or C++.
1460 The compiler can accept several base standards, such as @samp{c89} or
1461 @samp{c++98}, and GNU dialects of those standards, such as
1462 @samp{gnu89} or @samp{gnu++98}. By specifying a base standard, the
1463 compiler will accept all programs following that standard and those
1464 using GNU extensions that do not contradict it. For example,
1465 @samp{-std=c89} turns off certain features of GCC that are
1466 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1467 keywords, but not other GNU extensions that do not have a meaning in
1468 ISO C90, such as omitting the middle term of a @code{?:}
1469 expression. On the other hand, by specifying a GNU dialect of a
1470 standard, all features the compiler support are enabled, even when
1471 those features change the meaning of the base standard and some
1472 strict-conforming programs may be rejected. The particular standard
1473 is used by @option{-pedantic} to identify which features are GNU
1474 extensions given that version of the standard. For example
1475 @samp{-std=gnu89 -pedantic} would warn about C++ style @samp{//}
1476 comments, while @samp{-std=gnu99 -pedantic} would not.
1478 A value for this option must be provided; possible values are
1483 Support all ISO C90 programs (certain GNU extensions that conflict
1484 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1486 @item iso9899:199409
1487 ISO C90 as modified in amendment 1.
1493 ISO C99. Note that this standard is not yet fully supported; see
1494 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1495 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1498 GNU dialect of ISO C90 (including some C99 features). This
1499 is the default for C code.
1503 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1504 this will become the default. The name @samp{gnu9x} is deprecated.
1507 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1511 GNU dialect of @option{-std=c++98}. This is the default for
1515 The working draft of the upcoming ISO C++0x standard. This option
1516 enables experimental features that are likely to be included in
1517 C++0x. The working draft is constantly changing, and any feature that is
1518 enabled by this flag may be removed from future versions of GCC if it is
1519 not part of the C++0x standard.
1522 GNU dialect of @option{-std=c++0x}. This option enables
1523 experimental features that may be removed in future versions of GCC.
1526 @item -fgnu89-inline
1527 @opindex fgnu89-inline
1528 The option @option{-fgnu89-inline} tells GCC to use the traditional
1529 GNU semantics for @code{inline} functions when in C99 mode.
1530 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1531 is accepted and ignored by GCC versions 4.1.3 up to but not including
1532 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1533 C99 mode. Using this option is roughly equivalent to adding the
1534 @code{gnu_inline} function attribute to all inline functions
1535 (@pxref{Function Attributes}).
1537 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1538 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1539 specifies the default behavior). This option was first supported in
1540 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1542 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1543 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1544 in effect for @code{inline} functions. @xref{Common Predefined
1545 Macros,,,cpp,The C Preprocessor}.
1547 @item -aux-info @var{filename}
1549 Output to the given filename prototyped declarations for all functions
1550 declared and/or defined in a translation unit, including those in header
1551 files. This option is silently ignored in any language other than C@.
1553 Besides declarations, the file indicates, in comments, the origin of
1554 each declaration (source file and line), whether the declaration was
1555 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1556 @samp{O} for old, respectively, in the first character after the line
1557 number and the colon), and whether it came from a declaration or a
1558 definition (@samp{C} or @samp{F}, respectively, in the following
1559 character). In the case of function definitions, a K&R-style list of
1560 arguments followed by their declarations is also provided, inside
1561 comments, after the declaration.
1565 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1566 keyword, so that code can use these words as identifiers. You can use
1567 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1568 instead. @option{-ansi} implies @option{-fno-asm}.
1570 In C++, this switch only affects the @code{typeof} keyword, since
1571 @code{asm} and @code{inline} are standard keywords. You may want to
1572 use the @option{-fno-gnu-keywords} flag instead, which has the same
1573 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1574 switch only affects the @code{asm} and @code{typeof} keywords, since
1575 @code{inline} is a standard keyword in ISO C99.
1578 @itemx -fno-builtin-@var{function}
1579 @opindex fno-builtin
1580 @cindex built-in functions
1581 Don't recognize built-in functions that do not begin with
1582 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1583 functions provided by GCC}, for details of the functions affected,
1584 including those which are not built-in functions when @option{-ansi} or
1585 @option{-std} options for strict ISO C conformance are used because they
1586 do not have an ISO standard meaning.
1588 GCC normally generates special code to handle certain built-in functions
1589 more efficiently; for instance, calls to @code{alloca} may become single
1590 instructions that adjust the stack directly, and calls to @code{memcpy}
1591 may become inline copy loops. The resulting code is often both smaller
1592 and faster, but since the function calls no longer appear as such, you
1593 cannot set a breakpoint on those calls, nor can you change the behavior
1594 of the functions by linking with a different library. In addition,
1595 when a function is recognized as a built-in function, GCC may use
1596 information about that function to warn about problems with calls to
1597 that function, or to generate more efficient code, even if the
1598 resulting code still contains calls to that function. For example,
1599 warnings are given with @option{-Wformat} for bad calls to
1600 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1601 known not to modify global memory.
1603 With the @option{-fno-builtin-@var{function}} option
1604 only the built-in function @var{function} is
1605 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1606 function is named that is not built-in in this version of GCC, this
1607 option is ignored. There is no corresponding
1608 @option{-fbuiltin-@var{function}} option; if you wish to enable
1609 built-in functions selectively when using @option{-fno-builtin} or
1610 @option{-ffreestanding}, you may define macros such as:
1613 #define abs(n) __builtin_abs ((n))
1614 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1619 @cindex hosted environment
1621 Assert that compilation takes place in a hosted environment. This implies
1622 @option{-fbuiltin}. A hosted environment is one in which the
1623 entire standard library is available, and in which @code{main} has a return
1624 type of @code{int}. Examples are nearly everything except a kernel.
1625 This is equivalent to @option{-fno-freestanding}.
1627 @item -ffreestanding
1628 @opindex ffreestanding
1629 @cindex hosted environment
1631 Assert that compilation takes place in a freestanding environment. This
1632 implies @option{-fno-builtin}. A freestanding environment
1633 is one in which the standard library may not exist, and program startup may
1634 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1635 This is equivalent to @option{-fno-hosted}.
1637 @xref{Standards,,Language Standards Supported by GCC}, for details of
1638 freestanding and hosted environments.
1642 @cindex openmp parallel
1643 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1644 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1645 compiler generates parallel code according to the OpenMP Application
1646 Program Interface v3.0 @w{@uref{http://www.openmp.org/}}. This option
1647 implies @option{-pthread}, and thus is only supported on targets that
1648 have support for @option{-pthread}.
1650 @item -fms-extensions
1651 @opindex fms-extensions
1652 Accept some non-standard constructs used in Microsoft header files.
1654 Some cases of unnamed fields in structures and unions are only
1655 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1656 fields within structs/unions}, for details.
1660 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1661 options for strict ISO C conformance) implies @option{-trigraphs}.
1663 @item -no-integrated-cpp
1664 @opindex no-integrated-cpp
1665 Performs a compilation in two passes: preprocessing and compiling. This
1666 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1667 @option{-B} option. The user supplied compilation step can then add in
1668 an additional preprocessing step after normal preprocessing but before
1669 compiling. The default is to use the integrated cpp (internal cpp)
1671 The semantics of this option will change if "cc1", "cc1plus", and
1672 "cc1obj" are merged.
1674 @cindex traditional C language
1675 @cindex C language, traditional
1677 @itemx -traditional-cpp
1678 @opindex traditional-cpp
1679 @opindex traditional
1680 Formerly, these options caused GCC to attempt to emulate a pre-standard
1681 C compiler. They are now only supported with the @option{-E} switch.
1682 The preprocessor continues to support a pre-standard mode. See the GNU
1683 CPP manual for details.
1685 @item -fcond-mismatch
1686 @opindex fcond-mismatch
1687 Allow conditional expressions with mismatched types in the second and
1688 third arguments. The value of such an expression is void. This option
1689 is not supported for C++.
1691 @item -flax-vector-conversions
1692 @opindex flax-vector-conversions
1693 Allow implicit conversions between vectors with differing numbers of
1694 elements and/or incompatible element types. This option should not be
1697 @item -funsigned-char
1698 @opindex funsigned-char
1699 Let the type @code{char} be unsigned, like @code{unsigned char}.
1701 Each kind of machine has a default for what @code{char} should
1702 be. It is either like @code{unsigned char} by default or like
1703 @code{signed char} by default.
1705 Ideally, a portable program should always use @code{signed char} or
1706 @code{unsigned char} when it depends on the signedness of an object.
1707 But many programs have been written to use plain @code{char} and
1708 expect it to be signed, or expect it to be unsigned, depending on the
1709 machines they were written for. This option, and its inverse, let you
1710 make such a program work with the opposite default.
1712 The type @code{char} is always a distinct type from each of
1713 @code{signed char} or @code{unsigned char}, even though its behavior
1714 is always just like one of those two.
1717 @opindex fsigned-char
1718 Let the type @code{char} be signed, like @code{signed char}.
1720 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1721 the negative form of @option{-funsigned-char}. Likewise, the option
1722 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1724 @item -fsigned-bitfields
1725 @itemx -funsigned-bitfields
1726 @itemx -fno-signed-bitfields
1727 @itemx -fno-unsigned-bitfields
1728 @opindex fsigned-bitfields
1729 @opindex funsigned-bitfields
1730 @opindex fno-signed-bitfields
1731 @opindex fno-unsigned-bitfields
1732 These options control whether a bit-field is signed or unsigned, when the
1733 declaration does not use either @code{signed} or @code{unsigned}. By
1734 default, such a bit-field is signed, because this is consistent: the
1735 basic integer types such as @code{int} are signed types.
1738 @node C++ Dialect Options
1739 @section Options Controlling C++ Dialect
1741 @cindex compiler options, C++
1742 @cindex C++ options, command line
1743 @cindex options, C++
1744 This section describes the command-line options that are only meaningful
1745 for C++ programs; but you can also use most of the GNU compiler options
1746 regardless of what language your program is in. For example, you
1747 might compile a file @code{firstClass.C} like this:
1750 g++ -g -frepo -O -c firstClass.C
1754 In this example, only @option{-frepo} is an option meant
1755 only for C++ programs; you can use the other options with any
1756 language supported by GCC@.
1758 Here is a list of options that are @emph{only} for compiling C++ programs:
1762 @item -fabi-version=@var{n}
1763 @opindex fabi-version
1764 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1765 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1766 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1767 the version that conforms most closely to the C++ ABI specification.
1768 Therefore, the ABI obtained using version 0 will change as ABI bugs
1771 The default is version 2.
1773 @item -fno-access-control
1774 @opindex fno-access-control
1775 Turn off all access checking. This switch is mainly useful for working
1776 around bugs in the access control code.
1780 Check that the pointer returned by @code{operator new} is non-null
1781 before attempting to modify the storage allocated. This check is
1782 normally unnecessary because the C++ standard specifies that
1783 @code{operator new} will only return @code{0} if it is declared
1784 @samp{throw()}, in which case the compiler will always check the
1785 return value even without this option. In all other cases, when
1786 @code{operator new} has a non-empty exception specification, memory
1787 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1788 @samp{new (nothrow)}.
1790 @item -fconserve-space
1791 @opindex fconserve-space
1792 Put uninitialized or runtime-initialized global variables into the
1793 common segment, as C does. This saves space in the executable at the
1794 cost of not diagnosing duplicate definitions. If you compile with this
1795 flag and your program mysteriously crashes after @code{main()} has
1796 completed, you may have an object that is being destroyed twice because
1797 two definitions were merged.
1799 This option is no longer useful on most targets, now that support has
1800 been added for putting variables into BSS without making them common.
1802 @item -fno-deduce-init-list
1803 @opindex fno-deduce-init-list
1804 Disable deduction of a template type parameter as
1805 std::initializer_list from a brace-enclosed initializer list, i.e.
1808 template <class T> auto forward(T t) -> decltype (realfn (t))
1815 forward(@{1,2@}); // call forward<std::initializer_list<int>>
1819 This option is present because this deduction is an extension to the
1820 current specification in the C++0x working draft, and there was
1821 some concern about potential overload resolution problems.
1823 @item -ffriend-injection
1824 @opindex ffriend-injection
1825 Inject friend functions into the enclosing namespace, so that they are
1826 visible outside the scope of the class in which they are declared.
1827 Friend functions were documented to work this way in the old Annotated
1828 C++ Reference Manual, and versions of G++ before 4.1 always worked
1829 that way. However, in ISO C++ a friend function which is not declared
1830 in an enclosing scope can only be found using argument dependent
1831 lookup. This option causes friends to be injected as they were in
1834 This option is for compatibility, and may be removed in a future
1837 @item -fno-elide-constructors
1838 @opindex fno-elide-constructors
1839 The C++ standard allows an implementation to omit creating a temporary
1840 which is only used to initialize another object of the same type.
1841 Specifying this option disables that optimization, and forces G++ to
1842 call the copy constructor in all cases.
1844 @item -fno-enforce-eh-specs
1845 @opindex fno-enforce-eh-specs
1846 Don't generate code to check for violation of exception specifications
1847 at runtime. This option violates the C++ standard, but may be useful
1848 for reducing code size in production builds, much like defining
1849 @samp{NDEBUG}. This does not give user code permission to throw
1850 exceptions in violation of the exception specifications; the compiler
1851 will still optimize based on the specifications, so throwing an
1852 unexpected exception will result in undefined behavior.
1855 @itemx -fno-for-scope
1857 @opindex fno-for-scope
1858 If @option{-ffor-scope} is specified, the scope of variables declared in
1859 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1860 as specified by the C++ standard.
1861 If @option{-fno-for-scope} is specified, the scope of variables declared in
1862 a @i{for-init-statement} extends to the end of the enclosing scope,
1863 as was the case in old versions of G++, and other (traditional)
1864 implementations of C++.
1866 The default if neither flag is given to follow the standard,
1867 but to allow and give a warning for old-style code that would
1868 otherwise be invalid, or have different behavior.
1870 @item -fno-gnu-keywords
1871 @opindex fno-gnu-keywords
1872 Do not recognize @code{typeof} as a keyword, so that code can use this
1873 word as an identifier. You can use the keyword @code{__typeof__} instead.
1874 @option{-ansi} implies @option{-fno-gnu-keywords}.
1876 @item -fno-implicit-templates
1877 @opindex fno-implicit-templates
1878 Never emit code for non-inline templates which are instantiated
1879 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1880 @xref{Template Instantiation}, for more information.
1882 @item -fno-implicit-inline-templates
1883 @opindex fno-implicit-inline-templates
1884 Don't emit code for implicit instantiations of inline templates, either.
1885 The default is to handle inlines differently so that compiles with and
1886 without optimization will need the same set of explicit instantiations.
1888 @item -fno-implement-inlines
1889 @opindex fno-implement-inlines
1890 To save space, do not emit out-of-line copies of inline functions
1891 controlled by @samp{#pragma implementation}. This will cause linker
1892 errors if these functions are not inlined everywhere they are called.
1894 @item -fms-extensions
1895 @opindex fms-extensions
1896 Disable pedantic warnings about constructs used in MFC, such as implicit
1897 int and getting a pointer to member function via non-standard syntax.
1899 @item -fno-nonansi-builtins
1900 @opindex fno-nonansi-builtins
1901 Disable built-in declarations of functions that are not mandated by
1902 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1903 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1905 @item -fno-operator-names
1906 @opindex fno-operator-names
1907 Do not treat the operator name keywords @code{and}, @code{bitand},
1908 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1909 synonyms as keywords.
1911 @item -fno-optional-diags
1912 @opindex fno-optional-diags
1913 Disable diagnostics that the standard says a compiler does not need to
1914 issue. Currently, the only such diagnostic issued by G++ is the one for
1915 a name having multiple meanings within a class.
1918 @opindex fpermissive
1919 Downgrade some diagnostics about nonconformant code from errors to
1920 warnings. Thus, using @option{-fpermissive} will allow some
1921 nonconforming code to compile.
1923 @item -fno-pretty-templates
1924 @opindex fno-pretty-templates
1925 When an error message refers to a specialization of a function
1926 template, the compiler will normally print the signature of the
1927 template followed by the template arguments and any typedefs or
1928 typenames in the signature (e.g. @code{void f(T) [with T = int]}
1929 rather than @code{void f(int)}) so that it's clear which template is
1930 involved. When an error message refers to a specialization of a class
1931 template, the compiler will omit any template arguments which match
1932 the default template arguments for that template. If either of these
1933 behaviors make it harder to understand the error message rather than
1934 easier, using @option{-fno-pretty-templates} will disable them.
1938 Enable automatic template instantiation at link time. This option also
1939 implies @option{-fno-implicit-templates}. @xref{Template
1940 Instantiation}, for more information.
1944 Disable generation of information about every class with virtual
1945 functions for use by the C++ runtime type identification features
1946 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1947 of the language, you can save some space by using this flag. Note that
1948 exception handling uses the same information, but it will generate it as
1949 needed. The @samp{dynamic_cast} operator can still be used for casts that
1950 do not require runtime type information, i.e.@: casts to @code{void *} or to
1951 unambiguous base classes.
1955 Emit statistics about front-end processing at the end of the compilation.
1956 This information is generally only useful to the G++ development team.
1958 @item -ftemplate-depth-@var{n}
1959 @opindex ftemplate-depth
1960 Set the maximum instantiation depth for template classes to @var{n}.
1961 A limit on the template instantiation depth is needed to detect
1962 endless recursions during template class instantiation. ANSI/ISO C++
1963 conforming programs must not rely on a maximum depth greater than 17
1964 (changed to 1024 in C++0x).
1966 @item -fno-threadsafe-statics
1967 @opindex fno-threadsafe-statics
1968 Do not emit the extra code to use the routines specified in the C++
1969 ABI for thread-safe initialization of local statics. You can use this
1970 option to reduce code size slightly in code that doesn't need to be
1973 @item -fuse-cxa-atexit
1974 @opindex fuse-cxa-atexit
1975 Register destructors for objects with static storage duration with the
1976 @code{__cxa_atexit} function rather than the @code{atexit} function.
1977 This option is required for fully standards-compliant handling of static
1978 destructors, but will only work if your C library supports
1979 @code{__cxa_atexit}.
1981 @item -fno-use-cxa-get-exception-ptr
1982 @opindex fno-use-cxa-get-exception-ptr
1983 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1984 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1985 if the runtime routine is not available.
1987 @item -fvisibility-inlines-hidden
1988 @opindex fvisibility-inlines-hidden
1989 This switch declares that the user does not attempt to compare
1990 pointers to inline methods where the addresses of the two functions
1991 were taken in different shared objects.
1993 The effect of this is that GCC may, effectively, mark inline methods with
1994 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1995 appear in the export table of a DSO and do not require a PLT indirection
1996 when used within the DSO@. Enabling this option can have a dramatic effect
1997 on load and link times of a DSO as it massively reduces the size of the
1998 dynamic export table when the library makes heavy use of templates.
2000 The behavior of this switch is not quite the same as marking the
2001 methods as hidden directly, because it does not affect static variables
2002 local to the function or cause the compiler to deduce that
2003 the function is defined in only one shared object.
2005 You may mark a method as having a visibility explicitly to negate the
2006 effect of the switch for that method. For example, if you do want to
2007 compare pointers to a particular inline method, you might mark it as
2008 having default visibility. Marking the enclosing class with explicit
2009 visibility will have no effect.
2011 Explicitly instantiated inline methods are unaffected by this option
2012 as their linkage might otherwise cross a shared library boundary.
2013 @xref{Template Instantiation}.
2015 @item -fvisibility-ms-compat
2016 @opindex fvisibility-ms-compat
2017 This flag attempts to use visibility settings to make GCC's C++
2018 linkage model compatible with that of Microsoft Visual Studio.
2020 The flag makes these changes to GCC's linkage model:
2024 It sets the default visibility to @code{hidden}, like
2025 @option{-fvisibility=hidden}.
2028 Types, but not their members, are not hidden by default.
2031 The One Definition Rule is relaxed for types without explicit
2032 visibility specifications which are defined in more than one different
2033 shared object: those declarations are permitted if they would have
2034 been permitted when this option was not used.
2037 In new code it is better to use @option{-fvisibility=hidden} and
2038 export those classes which are intended to be externally visible.
2039 Unfortunately it is possible for code to rely, perhaps accidentally,
2040 on the Visual Studio behavior.
2042 Among the consequences of these changes are that static data members
2043 of the same type with the same name but defined in different shared
2044 objects will be different, so changing one will not change the other;
2045 and that pointers to function members defined in different shared
2046 objects may not compare equal. When this flag is given, it is a
2047 violation of the ODR to define types with the same name differently.
2051 Do not use weak symbol support, even if it is provided by the linker.
2052 By default, G++ will use weak symbols if they are available. This
2053 option exists only for testing, and should not be used by end-users;
2054 it will result in inferior code and has no benefits. This option may
2055 be removed in a future release of G++.
2059 Do not search for header files in the standard directories specific to
2060 C++, but do still search the other standard directories. (This option
2061 is used when building the C++ library.)
2064 In addition, these optimization, warning, and code generation options
2065 have meanings only for C++ programs:
2068 @item -fno-default-inline
2069 @opindex fno-default-inline
2070 Do not assume @samp{inline} for functions defined inside a class scope.
2071 @xref{Optimize Options,,Options That Control Optimization}. Note that these
2072 functions will have linkage like inline functions; they just won't be
2075 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2078 Warn when G++ generates code that is probably not compatible with the
2079 vendor-neutral C++ ABI@. Although an effort has been made to warn about
2080 all such cases, there are probably some cases that are not warned about,
2081 even though G++ is generating incompatible code. There may also be
2082 cases where warnings are emitted even though the code that is generated
2085 You should rewrite your code to avoid these warnings if you are
2086 concerned about the fact that code generated by G++ may not be binary
2087 compatible with code generated by other compilers.
2089 The known incompatibilities at this point include:
2094 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
2095 pack data into the same byte as a base class. For example:
2098 struct A @{ virtual void f(); int f1 : 1; @};
2099 struct B : public A @{ int f2 : 1; @};
2103 In this case, G++ will place @code{B::f2} into the same byte
2104 as@code{A::f1}; other compilers will not. You can avoid this problem
2105 by explicitly padding @code{A} so that its size is a multiple of the
2106 byte size on your platform; that will cause G++ and other compilers to
2107 layout @code{B} identically.
2110 Incorrect handling of tail-padding for virtual bases. G++ does not use
2111 tail padding when laying out virtual bases. For example:
2114 struct A @{ virtual void f(); char c1; @};
2115 struct B @{ B(); char c2; @};
2116 struct C : public A, public virtual B @{@};
2120 In this case, G++ will not place @code{B} into the tail-padding for
2121 @code{A}; other compilers will. You can avoid this problem by
2122 explicitly padding @code{A} so that its size is a multiple of its
2123 alignment (ignoring virtual base classes); that will cause G++ and other
2124 compilers to layout @code{C} identically.
2127 Incorrect handling of bit-fields with declared widths greater than that
2128 of their underlying types, when the bit-fields appear in a union. For
2132 union U @{ int i : 4096; @};
2136 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2137 union too small by the number of bits in an @code{int}.
2140 Empty classes can be placed at incorrect offsets. For example:
2150 struct C : public B, public A @{@};
2154 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2155 it should be placed at offset zero. G++ mistakenly believes that the
2156 @code{A} data member of @code{B} is already at offset zero.
2159 Names of template functions whose types involve @code{typename} or
2160 template template parameters can be mangled incorrectly.
2163 template <typename Q>
2164 void f(typename Q::X) @{@}
2166 template <template <typename> class Q>
2167 void f(typename Q<int>::X) @{@}
2171 Instantiations of these templates may be mangled incorrectly.
2175 It also warns psABI related changes. The known psABI changes at this
2181 For SYSV/x86-64, when passing union with long double, it is changed to
2182 pass in memory as specified in psABI. For example:
2192 @code{union U} will always be passed in memory.
2196 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2197 @opindex Wctor-dtor-privacy
2198 @opindex Wno-ctor-dtor-privacy
2199 Warn when a class seems unusable because all the constructors or
2200 destructors in that class are private, and it has neither friends nor
2201 public static member functions.
2203 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2204 @opindex Wnon-virtual-dtor
2205 @opindex Wno-non-virtual-dtor
2206 Warn when a class has virtual functions and accessible non-virtual
2207 destructor, in which case it would be possible but unsafe to delete
2208 an instance of a derived class through a pointer to the base class.
2209 This warning is also enabled if -Weffc++ is specified.
2211 @item -Wreorder @r{(C++ and Objective-C++ only)}
2213 @opindex Wno-reorder
2214 @cindex reordering, warning
2215 @cindex warning for reordering of member initializers
2216 Warn when the order of member initializers given in the code does not
2217 match the order in which they must be executed. For instance:
2223 A(): j (0), i (1) @{ @}
2227 The compiler will rearrange the member initializers for @samp{i}
2228 and @samp{j} to match the declaration order of the members, emitting
2229 a warning to that effect. This warning is enabled by @option{-Wall}.
2232 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2235 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2238 Warn about violations of the following style guidelines from Scott Meyers'
2239 @cite{Effective C++} book:
2243 Item 11: Define a copy constructor and an assignment operator for classes
2244 with dynamically allocated memory.
2247 Item 12: Prefer initialization to assignment in constructors.
2250 Item 14: Make destructors virtual in base classes.
2253 Item 15: Have @code{operator=} return a reference to @code{*this}.
2256 Item 23: Don't try to return a reference when you must return an object.
2260 Also warn about violations of the following style guidelines from
2261 Scott Meyers' @cite{More Effective C++} book:
2265 Item 6: Distinguish between prefix and postfix forms of increment and
2266 decrement operators.
2269 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2273 When selecting this option, be aware that the standard library
2274 headers do not obey all of these guidelines; use @samp{grep -v}
2275 to filter out those warnings.
2277 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2278 @opindex Wstrict-null-sentinel
2279 @opindex Wno-strict-null-sentinel
2280 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2281 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2282 to @code{__null}. Although it is a null pointer constant not a null pointer,
2283 it is guaranteed to be of the same size as a pointer. But this use is
2284 not portable across different compilers.
2286 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2287 @opindex Wno-non-template-friend
2288 @opindex Wnon-template-friend
2289 Disable warnings when non-templatized friend functions are declared
2290 within a template. Since the advent of explicit template specification
2291 support in G++, if the name of the friend is an unqualified-id (i.e.,
2292 @samp{friend foo(int)}), the C++ language specification demands that the
2293 friend declare or define an ordinary, nontemplate function. (Section
2294 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2295 could be interpreted as a particular specialization of a templatized
2296 function. Because this non-conforming behavior is no longer the default
2297 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2298 check existing code for potential trouble spots and is on by default.
2299 This new compiler behavior can be turned off with
2300 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2301 but disables the helpful warning.
2303 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2304 @opindex Wold-style-cast
2305 @opindex Wno-old-style-cast
2306 Warn if an old-style (C-style) cast to a non-void type is used within
2307 a C++ program. The new-style casts (@samp{dynamic_cast},
2308 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2309 less vulnerable to unintended effects and much easier to search for.
2311 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2312 @opindex Woverloaded-virtual
2313 @opindex Wno-overloaded-virtual
2314 @cindex overloaded virtual fn, warning
2315 @cindex warning for overloaded virtual fn
2316 Warn when a function declaration hides virtual functions from a
2317 base class. For example, in:
2324 struct B: public A @{
2329 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2337 will fail to compile.
2339 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2340 @opindex Wno-pmf-conversions
2341 @opindex Wpmf-conversions
2342 Disable the diagnostic for converting a bound pointer to member function
2345 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2346 @opindex Wsign-promo
2347 @opindex Wno-sign-promo
2348 Warn when overload resolution chooses a promotion from unsigned or
2349 enumerated type to a signed type, over a conversion to an unsigned type of
2350 the same size. Previous versions of G++ would try to preserve
2351 unsignedness, but the standard mandates the current behavior.
2356 A& operator = (int);
2366 In this example, G++ will synthesize a default @samp{A& operator =
2367 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2370 @node Objective-C and Objective-C++ Dialect Options
2371 @section Options Controlling Objective-C and Objective-C++ Dialects
2373 @cindex compiler options, Objective-C and Objective-C++
2374 @cindex Objective-C and Objective-C++ options, command line
2375 @cindex options, Objective-C and Objective-C++
2376 (NOTE: This manual does not describe the Objective-C and Objective-C++
2377 languages themselves. See @xref{Standards,,Language Standards
2378 Supported by GCC}, for references.)
2380 This section describes the command-line options that are only meaningful
2381 for Objective-C and Objective-C++ programs, but you can also use most of
2382 the language-independent GNU compiler options.
2383 For example, you might compile a file @code{some_class.m} like this:
2386 gcc -g -fgnu-runtime -O -c some_class.m
2390 In this example, @option{-fgnu-runtime} is an option meant only for
2391 Objective-C and Objective-C++ programs; you can use the other options with
2392 any language supported by GCC@.
2394 Note that since Objective-C is an extension of the C language, Objective-C
2395 compilations may also use options specific to the C front-end (e.g.,
2396 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2397 C++-specific options (e.g., @option{-Wabi}).
2399 Here is a list of options that are @emph{only} for compiling Objective-C
2400 and Objective-C++ programs:
2403 @item -fconstant-string-class=@var{class-name}
2404 @opindex fconstant-string-class
2405 Use @var{class-name} as the name of the class to instantiate for each
2406 literal string specified with the syntax @code{@@"@dots{}"}. The default
2407 class name is @code{NXConstantString} if the GNU runtime is being used, and
2408 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2409 @option{-fconstant-cfstrings} option, if also present, will override the
2410 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2411 to be laid out as constant CoreFoundation strings.
2414 @opindex fgnu-runtime
2415 Generate object code compatible with the standard GNU Objective-C
2416 runtime. This is the default for most types of systems.
2418 @item -fnext-runtime
2419 @opindex fnext-runtime
2420 Generate output compatible with the NeXT runtime. This is the default
2421 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2422 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2425 @item -fno-nil-receivers
2426 @opindex fno-nil-receivers
2427 Assume that all Objective-C message dispatches (e.g.,
2428 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2429 is not @code{nil}. This allows for more efficient entry points in the runtime
2430 to be used. Currently, this option is only available in conjunction with
2431 the NeXT runtime on Mac OS X 10.3 and later.
2433 @item -fobjc-call-cxx-cdtors
2434 @opindex fobjc-call-cxx-cdtors
2435 For each Objective-C class, check if any of its instance variables is a
2436 C++ object with a non-trivial default constructor. If so, synthesize a
2437 special @code{- (id) .cxx_construct} instance method that will run
2438 non-trivial default constructors on any such instance variables, in order,
2439 and then return @code{self}. Similarly, check if any instance variable
2440 is a C++ object with a non-trivial destructor, and if so, synthesize a
2441 special @code{- (void) .cxx_destruct} method that will run
2442 all such default destructors, in reverse order.
2444 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2445 thusly generated will only operate on instance variables declared in the
2446 current Objective-C class, and not those inherited from superclasses. It
2447 is the responsibility of the Objective-C runtime to invoke all such methods
2448 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2449 will be invoked by the runtime immediately after a new object
2450 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2451 be invoked immediately before the runtime deallocates an object instance.
2453 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2454 support for invoking the @code{- (id) .cxx_construct} and
2455 @code{- (void) .cxx_destruct} methods.
2457 @item -fobjc-direct-dispatch
2458 @opindex fobjc-direct-dispatch
2459 Allow fast jumps to the message dispatcher. On Darwin this is
2460 accomplished via the comm page.
2462 @item -fobjc-exceptions
2463 @opindex fobjc-exceptions
2464 Enable syntactic support for structured exception handling in Objective-C,
2465 similar to what is offered by C++ and Java. This option is
2466 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2475 @@catch (AnObjCClass *exc) @{
2482 @@catch (AnotherClass *exc) @{
2485 @@catch (id allOthers) @{
2495 The @code{@@throw} statement may appear anywhere in an Objective-C or
2496 Objective-C++ program; when used inside of a @code{@@catch} block, the
2497 @code{@@throw} may appear without an argument (as shown above), in which case
2498 the object caught by the @code{@@catch} will be rethrown.
2500 Note that only (pointers to) Objective-C objects may be thrown and
2501 caught using this scheme. When an object is thrown, it will be caught
2502 by the nearest @code{@@catch} clause capable of handling objects of that type,
2503 analogously to how @code{catch} blocks work in C++ and Java. A
2504 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2505 any and all Objective-C exceptions not caught by previous @code{@@catch}
2508 The @code{@@finally} clause, if present, will be executed upon exit from the
2509 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2510 regardless of whether any exceptions are thrown, caught or rethrown
2511 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2512 of the @code{finally} clause in Java.
2514 There are several caveats to using the new exception mechanism:
2518 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2519 idioms provided by the @code{NSException} class, the new
2520 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2521 systems, due to additional functionality needed in the (NeXT) Objective-C
2525 As mentioned above, the new exceptions do not support handling
2526 types other than Objective-C objects. Furthermore, when used from
2527 Objective-C++, the Objective-C exception model does not interoperate with C++
2528 exceptions at this time. This means you cannot @code{@@throw} an exception
2529 from Objective-C and @code{catch} it in C++, or vice versa
2530 (i.e., @code{throw @dots{} @@catch}).
2533 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2534 blocks for thread-safe execution:
2537 @@synchronized (ObjCClass *guard) @{
2542 Upon entering the @code{@@synchronized} block, a thread of execution shall
2543 first check whether a lock has been placed on the corresponding @code{guard}
2544 object by another thread. If it has, the current thread shall wait until
2545 the other thread relinquishes its lock. Once @code{guard} becomes available,
2546 the current thread will place its own lock on it, execute the code contained in
2547 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2548 making @code{guard} available to other threads).
2550 Unlike Java, Objective-C does not allow for entire methods to be marked
2551 @code{@@synchronized}. Note that throwing exceptions out of
2552 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2553 to be unlocked properly.
2557 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2559 @item -freplace-objc-classes
2560 @opindex freplace-objc-classes
2561 Emit a special marker instructing @command{ld(1)} not to statically link in
2562 the resulting object file, and allow @command{dyld(1)} to load it in at
2563 run time instead. This is used in conjunction with the Fix-and-Continue
2564 debugging mode, where the object file in question may be recompiled and
2565 dynamically reloaded in the course of program execution, without the need
2566 to restart the program itself. Currently, Fix-and-Continue functionality
2567 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2572 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2573 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2574 compile time) with static class references that get initialized at load time,
2575 which improves run-time performance. Specifying the @option{-fzero-link} flag
2576 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2577 to be retained. This is useful in Zero-Link debugging mode, since it allows
2578 for individual class implementations to be modified during program execution.
2582 Dump interface declarations for all classes seen in the source file to a
2583 file named @file{@var{sourcename}.decl}.
2585 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2586 @opindex Wassign-intercept
2587 @opindex Wno-assign-intercept
2588 Warn whenever an Objective-C assignment is being intercepted by the
2591 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2592 @opindex Wno-protocol
2594 If a class is declared to implement a protocol, a warning is issued for
2595 every method in the protocol that is not implemented by the class. The
2596 default behavior is to issue a warning for every method not explicitly
2597 implemented in the class, even if a method implementation is inherited
2598 from the superclass. If you use the @option{-Wno-protocol} option, then
2599 methods inherited from the superclass are considered to be implemented,
2600 and no warning is issued for them.
2602 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2604 @opindex Wno-selector
2605 Warn if multiple methods of different types for the same selector are
2606 found during compilation. The check is performed on the list of methods
2607 in the final stage of compilation. Additionally, a check is performed
2608 for each selector appearing in a @code{@@selector(@dots{})}
2609 expression, and a corresponding method for that selector has been found
2610 during compilation. Because these checks scan the method table only at
2611 the end of compilation, these warnings are not produced if the final
2612 stage of compilation is not reached, for example because an error is
2613 found during compilation, or because the @option{-fsyntax-only} option is
2616 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2617 @opindex Wstrict-selector-match
2618 @opindex Wno-strict-selector-match
2619 Warn if multiple methods with differing argument and/or return types are
2620 found for a given selector when attempting to send a message using this
2621 selector to a receiver of type @code{id} or @code{Class}. When this flag
2622 is off (which is the default behavior), the compiler will omit such warnings
2623 if any differences found are confined to types which share the same size
2626 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2627 @opindex Wundeclared-selector
2628 @opindex Wno-undeclared-selector
2629 Warn if a @code{@@selector(@dots{})} expression referring to an
2630 undeclared selector is found. A selector is considered undeclared if no
2631 method with that name has been declared before the
2632 @code{@@selector(@dots{})} expression, either explicitly in an
2633 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2634 an @code{@@implementation} section. This option always performs its
2635 checks as soon as a @code{@@selector(@dots{})} expression is found,
2636 while @option{-Wselector} only performs its checks in the final stage of
2637 compilation. This also enforces the coding style convention
2638 that methods and selectors must be declared before being used.
2640 @item -print-objc-runtime-info
2641 @opindex print-objc-runtime-info
2642 Generate C header describing the largest structure that is passed by
2647 @node Language Independent Options
2648 @section Options to Control Diagnostic Messages Formatting
2649 @cindex options to control diagnostics formatting
2650 @cindex diagnostic messages
2651 @cindex message formatting
2653 Traditionally, diagnostic messages have been formatted irrespective of
2654 the output device's aspect (e.g.@: its width, @dots{}). The options described
2655 below can be used to control the diagnostic messages formatting
2656 algorithm, e.g.@: how many characters per line, how often source location
2657 information should be reported. Right now, only the C++ front end can
2658 honor these options. However it is expected, in the near future, that
2659 the remaining front ends would be able to digest them correctly.
2662 @item -fmessage-length=@var{n}
2663 @opindex fmessage-length
2664 Try to format error messages so that they fit on lines of about @var{n}
2665 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2666 the front ends supported by GCC@. If @var{n} is zero, then no
2667 line-wrapping will be done; each error message will appear on a single
2670 @opindex fdiagnostics-show-location
2671 @item -fdiagnostics-show-location=once
2672 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2673 reporter to emit @emph{once} source location information; that is, in
2674 case the message is too long to fit on a single physical line and has to
2675 be wrapped, the source location won't be emitted (as prefix) again,
2676 over and over, in subsequent continuation lines. This is the default
2679 @item -fdiagnostics-show-location=every-line
2680 Only meaningful in line-wrapping mode. Instructs the diagnostic
2681 messages reporter to emit the same source location information (as
2682 prefix) for physical lines that result from the process of breaking
2683 a message which is too long to fit on a single line.
2685 @item -fdiagnostics-show-option
2686 @opindex fdiagnostics-show-option
2687 This option instructs the diagnostic machinery to add text to each
2688 diagnostic emitted, which indicates which command line option directly
2689 controls that diagnostic, when such an option is known to the
2690 diagnostic machinery.
2692 @item -Wcoverage-mismatch
2693 @opindex Wcoverage-mismatch
2694 Warn if feedback profiles do not match when using the
2695 @option{-fprofile-use} option.
2696 If a source file was changed between @option{-fprofile-gen} and
2697 @option{-fprofile-use}, the files with the profile feedback can fail
2698 to match the source file and GCC can not use the profile feedback
2699 information. By default, GCC emits an error message in this case.
2700 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2701 error. GCC does not use appropriate feedback profiles, so using this
2702 option can result in poorly optimized code. This option is useful
2703 only in the case of very minor changes such as bug fixes to an
2708 @node Warning Options
2709 @section Options to Request or Suppress Warnings
2710 @cindex options to control warnings
2711 @cindex warning messages
2712 @cindex messages, warning
2713 @cindex suppressing warnings
2715 Warnings are diagnostic messages that report constructions which
2716 are not inherently erroneous but which are risky or suggest there
2717 may have been an error.
2719 The following language-independent options do not enable specific
2720 warnings but control the kinds of diagnostics produced by GCC.
2723 @cindex syntax checking
2725 @opindex fsyntax-only
2726 Check the code for syntax errors, but don't do anything beyond that.
2730 Inhibit all warning messages.
2735 Make all warnings into errors.
2740 Make the specified warning into an error. The specifier for a warning
2741 is appended, for example @option{-Werror=switch} turns the warnings
2742 controlled by @option{-Wswitch} into errors. This switch takes a
2743 negative form, to be used to negate @option{-Werror} for specific
2744 warnings, for example @option{-Wno-error=switch} makes
2745 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2746 is in effect. You can use the @option{-fdiagnostics-show-option}
2747 option to have each controllable warning amended with the option which
2748 controls it, to determine what to use with this option.
2750 Note that specifying @option{-Werror=}@var{foo} automatically implies
2751 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2754 @item -Wfatal-errors
2755 @opindex Wfatal-errors
2756 @opindex Wno-fatal-errors
2757 This option causes the compiler to abort compilation on the first error
2758 occurred rather than trying to keep going and printing further error
2763 You can request many specific warnings with options beginning
2764 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2765 implicit declarations. Each of these specific warning options also
2766 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2767 example, @option{-Wno-implicit}. This manual lists only one of the
2768 two forms, whichever is not the default. For further,
2769 language-specific options also refer to @ref{C++ Dialect Options} and
2770 @ref{Objective-C and Objective-C++ Dialect Options}.
2775 Issue all the warnings demanded by strict ISO C and ISO C++;
2776 reject all programs that use forbidden extensions, and some other
2777 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2778 version of the ISO C standard specified by any @option{-std} option used.
2780 Valid ISO C and ISO C++ programs should compile properly with or without
2781 this option (though a rare few will require @option{-ansi} or a
2782 @option{-std} option specifying the required version of ISO C)@. However,
2783 without this option, certain GNU extensions and traditional C and C++
2784 features are supported as well. With this option, they are rejected.
2786 @option{-pedantic} does not cause warning messages for use of the
2787 alternate keywords whose names begin and end with @samp{__}. Pedantic
2788 warnings are also disabled in the expression that follows
2789 @code{__extension__}. However, only system header files should use
2790 these escape routes; application programs should avoid them.
2791 @xref{Alternate Keywords}.
2793 Some users try to use @option{-pedantic} to check programs for strict ISO
2794 C conformance. They soon find that it does not do quite what they want:
2795 it finds some non-ISO practices, but not all---only those for which
2796 ISO C @emph{requires} a diagnostic, and some others for which
2797 diagnostics have been added.
2799 A feature to report any failure to conform to ISO C might be useful in
2800 some instances, but would require considerable additional work and would
2801 be quite different from @option{-pedantic}. We don't have plans to
2802 support such a feature in the near future.
2804 Where the standard specified with @option{-std} represents a GNU
2805 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2806 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2807 extended dialect is based. Warnings from @option{-pedantic} are given
2808 where they are required by the base standard. (It would not make sense
2809 for such warnings to be given only for features not in the specified GNU
2810 C dialect, since by definition the GNU dialects of C include all
2811 features the compiler supports with the given option, and there would be
2812 nothing to warn about.)
2814 @item -pedantic-errors
2815 @opindex pedantic-errors
2816 Like @option{-pedantic}, except that errors are produced rather than
2822 This enables all the warnings about constructions that some users
2823 consider questionable, and that are easy to avoid (or modify to
2824 prevent the warning), even in conjunction with macros. This also
2825 enables some language-specific warnings described in @ref{C++ Dialect
2826 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2828 @option{-Wall} turns on the following warning flags:
2830 @gccoptlist{-Waddress @gol
2831 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2833 -Wchar-subscripts @gol
2834 -Wenum-compare @r{(in C/Objc; this is on by default in C++)} @gol
2836 -Wimplicit-function-declaration @gol
2839 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2840 -Wmissing-braces @gol
2846 -Wsequence-point @gol
2847 -Wsign-compare @r{(only in C++)} @gol
2848 -Wstrict-aliasing @gol
2849 -Wstrict-overflow=1 @gol
2852 -Wuninitialized @gol
2853 -Wunknown-pragmas @gol
2854 -Wunused-function @gol
2857 -Wunused-variable @gol
2858 -Wvolatile-register-var @gol
2861 Note that some warning flags are not implied by @option{-Wall}. Some of
2862 them warn about constructions that users generally do not consider
2863 questionable, but which occasionally you might wish to check for;
2864 others warn about constructions that are necessary or hard to avoid in
2865 some cases, and there is no simple way to modify the code to suppress
2866 the warning. Some of them are enabled by @option{-Wextra} but many of
2867 them must be enabled individually.
2873 This enables some extra warning flags that are not enabled by
2874 @option{-Wall}. (This option used to be called @option{-W}. The older
2875 name is still supported, but the newer name is more descriptive.)
2877 @gccoptlist{-Wclobbered @gol
2879 -Wignored-qualifiers @gol
2880 -Wmissing-field-initializers @gol
2881 -Wmissing-parameter-type @r{(C only)} @gol
2882 -Wold-style-declaration @r{(C only)} @gol
2883 -Woverride-init @gol
2886 -Wuninitialized @gol
2887 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2890 The option @option{-Wextra} also prints warning messages for the
2896 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2897 @samp{>}, or @samp{>=}.
2900 (C++ only) An enumerator and a non-enumerator both appear in a
2901 conditional expression.
2904 (C++ only) Ambiguous virtual bases.
2907 (C++ only) Subscripting an array which has been declared @samp{register}.
2910 (C++ only) Taking the address of a variable which has been declared
2914 (C++ only) A base class is not initialized in a derived class' copy
2919 @item -Wchar-subscripts
2920 @opindex Wchar-subscripts
2921 @opindex Wno-char-subscripts
2922 Warn if an array subscript has type @code{char}. This is a common cause
2923 of error, as programmers often forget that this type is signed on some
2925 This warning is enabled by @option{-Wall}.
2929 @opindex Wno-comment
2930 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2931 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2932 This warning is enabled by @option{-Wall}.
2937 @opindex ffreestanding
2938 @opindex fno-builtin
2939 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2940 the arguments supplied have types appropriate to the format string
2941 specified, and that the conversions specified in the format string make
2942 sense. This includes standard functions, and others specified by format
2943 attributes (@pxref{Function Attributes}), in the @code{printf},
2944 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2945 not in the C standard) families (or other target-specific families).
2946 Which functions are checked without format attributes having been
2947 specified depends on the standard version selected, and such checks of
2948 functions without the attribute specified are disabled by
2949 @option{-ffreestanding} or @option{-fno-builtin}.
2951 The formats are checked against the format features supported by GNU
2952 libc version 2.2. These include all ISO C90 and C99 features, as well
2953 as features from the Single Unix Specification and some BSD and GNU
2954 extensions. Other library implementations may not support all these
2955 features; GCC does not support warning about features that go beyond a
2956 particular library's limitations. However, if @option{-pedantic} is used
2957 with @option{-Wformat}, warnings will be given about format features not
2958 in the selected standard version (but not for @code{strfmon} formats,
2959 since those are not in any version of the C standard). @xref{C Dialect
2960 Options,,Options Controlling C Dialect}.
2962 Since @option{-Wformat} also checks for null format arguments for
2963 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2965 @option{-Wformat} is included in @option{-Wall}. For more control over some
2966 aspects of format checking, the options @option{-Wformat-y2k},
2967 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2968 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2969 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2972 @opindex Wformat-y2k
2973 @opindex Wno-format-y2k
2974 If @option{-Wformat} is specified, also warn about @code{strftime}
2975 formats which may yield only a two-digit year.
2977 @item -Wno-format-contains-nul
2978 @opindex Wno-format-contains-nul
2979 @opindex Wformat-contains-nul
2980 If @option{-Wformat} is specified, do not warn about format strings that
2983 @item -Wno-format-extra-args
2984 @opindex Wno-format-extra-args
2985 @opindex Wformat-extra-args
2986 If @option{-Wformat} is specified, do not warn about excess arguments to a
2987 @code{printf} or @code{scanf} format function. The C standard specifies
2988 that such arguments are ignored.
2990 Where the unused arguments lie between used arguments that are
2991 specified with @samp{$} operand number specifications, normally
2992 warnings are still given, since the implementation could not know what
2993 type to pass to @code{va_arg} to skip the unused arguments. However,
2994 in the case of @code{scanf} formats, this option will suppress the
2995 warning if the unused arguments are all pointers, since the Single
2996 Unix Specification says that such unused arguments are allowed.
2998 @item -Wno-format-zero-length @r{(C and Objective-C only)}
2999 @opindex Wno-format-zero-length
3000 @opindex Wformat-zero-length
3001 If @option{-Wformat} is specified, do not warn about zero-length formats.
3002 The C standard specifies that zero-length formats are allowed.
3004 @item -Wformat-nonliteral
3005 @opindex Wformat-nonliteral
3006 @opindex Wno-format-nonliteral
3007 If @option{-Wformat} is specified, also warn if the format string is not a
3008 string literal and so cannot be checked, unless the format function
3009 takes its format arguments as a @code{va_list}.
3011 @item -Wformat-security
3012 @opindex Wformat-security
3013 @opindex Wno-format-security
3014 If @option{-Wformat} is specified, also warn about uses of format
3015 functions that represent possible security problems. At present, this
3016 warns about calls to @code{printf} and @code{scanf} functions where the
3017 format string is not a string literal and there are no format arguments,
3018 as in @code{printf (foo);}. This may be a security hole if the format
3019 string came from untrusted input and contains @samp{%n}. (This is
3020 currently a subset of what @option{-Wformat-nonliteral} warns about, but
3021 in future warnings may be added to @option{-Wformat-security} that are not
3022 included in @option{-Wformat-nonliteral}.)
3026 @opindex Wno-format=2
3027 Enable @option{-Wformat} plus format checks not included in
3028 @option{-Wformat}. Currently equivalent to @samp{-Wformat
3029 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
3031 @item -Wnonnull @r{(C and Objective-C only)}
3033 @opindex Wno-nonnull
3034 Warn about passing a null pointer for arguments marked as
3035 requiring a non-null value by the @code{nonnull} function attribute.
3037 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
3038 can be disabled with the @option{-Wno-nonnull} option.
3040 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
3042 @opindex Wno-init-self
3043 Warn about uninitialized variables which are initialized with themselves.
3044 Note this option can only be used with the @option{-Wuninitialized} option.
3046 For example, GCC will warn about @code{i} being uninitialized in the
3047 following snippet only when @option{-Winit-self} has been specified:
3058 @item -Wimplicit-int @r{(C and Objective-C only)}
3059 @opindex Wimplicit-int
3060 @opindex Wno-implicit-int
3061 Warn when a declaration does not specify a type.
3062 This warning is enabled by @option{-Wall}.
3064 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
3065 @opindex Wimplicit-function-declaration
3066 @opindex Wno-implicit-function-declaration
3067 Give a warning whenever a function is used before being declared. In
3068 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
3069 enabled by default and it is made into an error by
3070 @option{-pedantic-errors}. This warning is also enabled by
3075 @opindex Wno-implicit
3076 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
3077 This warning is enabled by @option{-Wall}.
3079 @item -Wignored-qualifiers @r{(C and C++ only)}
3080 @opindex Wignored-qualifiers
3081 @opindex Wno-ignored-qualifiers
3082 Warn if the return type of a function has a type qualifier
3083 such as @code{const}. For ISO C such a type qualifier has no effect,
3084 since the value returned by a function is not an lvalue.
3085 For C++, the warning is only emitted for scalar types or @code{void}.
3086 ISO C prohibits qualified @code{void} return types on function
3087 definitions, so such return types always receive a warning
3088 even without this option.
3090 This warning is also enabled by @option{-Wextra}.
3095 Warn if the type of @samp{main} is suspicious. @samp{main} should be
3096 a function with external linkage, returning int, taking either zero
3097 arguments, two, or three arguments of appropriate types. This warning
3098 is enabled by default in C++ and is enabled by either @option{-Wall}
3099 or @option{-pedantic}.
3101 @item -Wmissing-braces
3102 @opindex Wmissing-braces
3103 @opindex Wno-missing-braces
3104 Warn if an aggregate or union initializer is not fully bracketed. In
3105 the following example, the initializer for @samp{a} is not fully
3106 bracketed, but that for @samp{b} is fully bracketed.
3109 int a[2][2] = @{ 0, 1, 2, 3 @};
3110 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
3113 This warning is enabled by @option{-Wall}.
3115 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
3116 @opindex Wmissing-include-dirs
3117 @opindex Wno-missing-include-dirs
3118 Warn if a user-supplied include directory does not exist.
3121 @opindex Wparentheses
3122 @opindex Wno-parentheses
3123 Warn if parentheses are omitted in certain contexts, such
3124 as when there is an assignment in a context where a truth value
3125 is expected, or when operators are nested whose precedence people
3126 often get confused about.
3128 Also warn if a comparison like @samp{x<=y<=z} appears; this is
3129 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
3130 interpretation from that of ordinary mathematical notation.
3132 Also warn about constructions where there may be confusion to which
3133 @code{if} statement an @code{else} branch belongs. Here is an example of
3148 In C/C++, every @code{else} branch belongs to the innermost possible
3149 @code{if} statement, which in this example is @code{if (b)}. This is
3150 often not what the programmer expected, as illustrated in the above
3151 example by indentation the programmer chose. When there is the
3152 potential for this confusion, GCC will issue a warning when this flag
3153 is specified. To eliminate the warning, add explicit braces around
3154 the innermost @code{if} statement so there is no way the @code{else}
3155 could belong to the enclosing @code{if}. The resulting code would
3172 This warning is enabled by @option{-Wall}.
3174 @item -Wsequence-point
3175 @opindex Wsequence-point
3176 @opindex Wno-sequence-point
3177 Warn about code that may have undefined semantics because of violations
3178 of sequence point rules in the C and C++ standards.
3180 The C and C++ standards defines the order in which expressions in a C/C++
3181 program are evaluated in terms of @dfn{sequence points}, which represent
3182 a partial ordering between the execution of parts of the program: those
3183 executed before the sequence point, and those executed after it. These
3184 occur after the evaluation of a full expression (one which is not part
3185 of a larger expression), after the evaluation of the first operand of a
3186 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3187 function is called (but after the evaluation of its arguments and the
3188 expression denoting the called function), and in certain other places.
3189 Other than as expressed by the sequence point rules, the order of
3190 evaluation of subexpressions of an expression is not specified. All
3191 these rules describe only a partial order rather than a total order,
3192 since, for example, if two functions are called within one expression
3193 with no sequence point between them, the order in which the functions
3194 are called is not specified. However, the standards committee have
3195 ruled that function calls do not overlap.
3197 It is not specified when between sequence points modifications to the
3198 values of objects take effect. Programs whose behavior depends on this
3199 have undefined behavior; the C and C++ standards specify that ``Between
3200 the previous and next sequence point an object shall have its stored
3201 value modified at most once by the evaluation of an expression.
3202 Furthermore, the prior value shall be read only to determine the value
3203 to be stored.''. If a program breaks these rules, the results on any
3204 particular implementation are entirely unpredictable.
3206 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3207 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3208 diagnosed by this option, and it may give an occasional false positive
3209 result, but in general it has been found fairly effective at detecting
3210 this sort of problem in programs.
3212 The standard is worded confusingly, therefore there is some debate
3213 over the precise meaning of the sequence point rules in subtle cases.
3214 Links to discussions of the problem, including proposed formal
3215 definitions, may be found on the GCC readings page, at
3216 @w{@uref{http://gcc.gnu.org/readings.html}}.
3218 This warning is enabled by @option{-Wall} for C and C++.
3221 @opindex Wreturn-type
3222 @opindex Wno-return-type
3223 Warn whenever a function is defined with a return-type that defaults
3224 to @code{int}. Also warn about any @code{return} statement with no
3225 return-value in a function whose return-type is not @code{void}
3226 (falling off the end of the function body is considered returning
3227 without a value), and about a @code{return} statement with an
3228 expression in a function whose return-type is @code{void}.
3230 For C++, a function without return type always produces a diagnostic
3231 message, even when @option{-Wno-return-type} is specified. The only
3232 exceptions are @samp{main} and functions defined in system headers.
3234 This warning is enabled by @option{-Wall}.
3239 Warn whenever a @code{switch} statement has an index of enumerated type
3240 and lacks a @code{case} for one or more of the named codes of that
3241 enumeration. (The presence of a @code{default} label prevents this
3242 warning.) @code{case} labels outside the enumeration range also
3243 provoke warnings when this option is used (even if there is a
3244 @code{default} label).
3245 This warning is enabled by @option{-Wall}.
3247 @item -Wswitch-default
3248 @opindex Wswitch-default
3249 @opindex Wno-switch-default
3250 Warn whenever a @code{switch} statement does not have a @code{default}
3254 @opindex Wswitch-enum
3255 @opindex Wno-switch-enum
3256 Warn whenever a @code{switch} statement has an index of enumerated type
3257 and lacks a @code{case} for one or more of the named codes of that
3258 enumeration. @code{case} labels outside the enumeration range also
3259 provoke warnings when this option is used. The only difference
3260 between @option{-Wswitch} and this option is that this option gives a
3261 warning about an omitted enumeration code even if there is a
3262 @code{default} label.
3264 @item -Wsync-nand @r{(C and C++ only)}
3266 @opindex Wno-sync-nand
3267 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
3268 built-in functions are used. These functions changed semantics in GCC 4.4.
3272 @opindex Wno-trigraphs
3273 Warn if any trigraphs are encountered that might change the meaning of
3274 the program (trigraphs within comments are not warned about).
3275 This warning is enabled by @option{-Wall}.
3277 @item -Wunused-function
3278 @opindex Wunused-function
3279 @opindex Wno-unused-function
3280 Warn whenever a static function is declared but not defined or a
3281 non-inline static function is unused.
3282 This warning is enabled by @option{-Wall}.
3284 @item -Wunused-label
3285 @opindex Wunused-label
3286 @opindex Wno-unused-label
3287 Warn whenever a label is declared but not used.
3288 This warning is enabled by @option{-Wall}.
3290 To suppress this warning use the @samp{unused} attribute
3291 (@pxref{Variable Attributes}).
3293 @item -Wunused-parameter
3294 @opindex Wunused-parameter
3295 @opindex Wno-unused-parameter
3296 Warn whenever a function parameter is unused aside from its declaration.
3298 To suppress this warning use the @samp{unused} attribute
3299 (@pxref{Variable Attributes}).
3301 @item -Wno-unused-result
3302 @opindex Wunused-result
3303 @opindex Wno-unused-result
3304 Do not warn if a caller of a function marked with attribute
3305 @code{warn_unused_result} (@pxref{Variable Attributes}) does not use
3306 its return value. The default is @option{-Wunused-result}.
3308 @item -Wunused-variable
3309 @opindex Wunused-variable
3310 @opindex Wno-unused-variable
3311 Warn whenever a local variable or non-constant static variable is unused
3312 aside from its declaration.
3313 This warning is enabled by @option{-Wall}.
3315 To suppress this warning use the @samp{unused} attribute
3316 (@pxref{Variable Attributes}).
3318 @item -Wunused-value
3319 @opindex Wunused-value
3320 @opindex Wno-unused-value
3321 Warn whenever a statement computes a result that is explicitly not
3322 used. To suppress this warning cast the unused expression to
3323 @samp{void}. This includes an expression-statement or the left-hand
3324 side of a comma expression that contains no side effects. For example,
3325 an expression such as @samp{x[i,j]} will cause a warning, while
3326 @samp{x[(void)i,j]} will not.
3328 This warning is enabled by @option{-Wall}.
3333 All the above @option{-Wunused} options combined.
3335 In order to get a warning about an unused function parameter, you must
3336 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3337 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3339 @item -Wuninitialized
3340 @opindex Wuninitialized
3341 @opindex Wno-uninitialized
3342 Warn if an automatic variable is used without first being initialized
3343 or if a variable may be clobbered by a @code{setjmp} call. In C++,
3344 warn if a non-static reference or non-static @samp{const} member
3345 appears in a class without constructors.
3347 If you want to warn about code which uses the uninitialized value of the
3348 variable in its own initializer, use the @option{-Winit-self} option.
3350 These warnings occur for individual uninitialized or clobbered
3351 elements of structure, union or array variables as well as for
3352 variables which are uninitialized or clobbered as a whole. They do
3353 not occur for variables or elements declared @code{volatile}. Because
3354 these warnings depend on optimization, the exact variables or elements
3355 for which there are warnings will depend on the precise optimization
3356 options and version of GCC used.
3358 Note that there may be no warning about a variable that is used only
3359 to compute a value that itself is never used, because such
3360 computations may be deleted by data flow analysis before the warnings
3363 These warnings are made optional because GCC is not smart
3364 enough to see all the reasons why the code might be correct
3365 despite appearing to have an error. Here is one example of how
3386 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3387 always initialized, but GCC doesn't know this. Here is
3388 another common case:
3393 if (change_y) save_y = y, y = new_y;
3395 if (change_y) y = save_y;
3400 This has no bug because @code{save_y} is used only if it is set.
3402 @cindex @code{longjmp} warnings
3403 This option also warns when a non-volatile automatic variable might be
3404 changed by a call to @code{longjmp}. These warnings as well are possible
3405 only in optimizing compilation.
3407 The compiler sees only the calls to @code{setjmp}. It cannot know
3408 where @code{longjmp} will be called; in fact, a signal handler could
3409 call it at any point in the code. As a result, you may get a warning
3410 even when there is in fact no problem because @code{longjmp} cannot
3411 in fact be called at the place which would cause a problem.
3413 Some spurious warnings can be avoided if you declare all the functions
3414 you use that never return as @code{noreturn}. @xref{Function
3417 This warning is enabled by @option{-Wall} or @option{-Wextra}.
3419 @item -Wunknown-pragmas
3420 @opindex Wunknown-pragmas
3421 @opindex Wno-unknown-pragmas
3422 @cindex warning for unknown pragmas
3423 @cindex unknown pragmas, warning
3424 @cindex pragmas, warning of unknown
3425 Warn when a #pragma directive is encountered which is not understood by
3426 GCC@. If this command line option is used, warnings will even be issued
3427 for unknown pragmas in system header files. This is not the case if
3428 the warnings were only enabled by the @option{-Wall} command line option.
3431 @opindex Wno-pragmas
3433 Do not warn about misuses of pragmas, such as incorrect parameters,
3434 invalid syntax, or conflicts between pragmas. See also
3435 @samp{-Wunknown-pragmas}.
3437 @item -Wstrict-aliasing
3438 @opindex Wstrict-aliasing
3439 @opindex Wno-strict-aliasing
3440 This option is only active when @option{-fstrict-aliasing} is active.
3441 It warns about code which might break the strict aliasing rules that the
3442 compiler is using for optimization. The warning does not catch all
3443 cases, but does attempt to catch the more common pitfalls. It is
3444 included in @option{-Wall}.
3445 It is equivalent to @option{-Wstrict-aliasing=3}
3447 @item -Wstrict-aliasing=n
3448 @opindex Wstrict-aliasing=n
3449 @opindex Wno-strict-aliasing=n
3450 This option is only active when @option{-fstrict-aliasing} is active.
3451 It warns about code which might break the strict aliasing rules that the
3452 compiler is using for optimization.
3453 Higher levels correspond to higher accuracy (fewer false positives).
3454 Higher levels also correspond to more effort, similar to the way -O works.
3455 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3458 Level 1: Most aggressive, quick, least accurate.
3459 Possibly useful when higher levels
3460 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3461 false negatives. However, it has many false positives.
3462 Warns for all pointer conversions between possibly incompatible types,
3463 even if never dereferenced. Runs in the frontend only.
3465 Level 2: Aggressive, quick, not too precise.
3466 May still have many false positives (not as many as level 1 though),
3467 and few false negatives (but possibly more than level 1).
3468 Unlike level 1, it only warns when an address is taken. Warns about
3469 incomplete types. Runs in the frontend only.
3471 Level 3 (default for @option{-Wstrict-aliasing}):
3472 Should have very few false positives and few false
3473 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3474 Takes care of the common punn+dereference pattern in the frontend:
3475 @code{*(int*)&some_float}.
3476 If optimization is enabled, it also runs in the backend, where it deals
3477 with multiple statement cases using flow-sensitive points-to information.
3478 Only warns when the converted pointer is dereferenced.
3479 Does not warn about incomplete types.
3481 @item -Wstrict-overflow
3482 @itemx -Wstrict-overflow=@var{n}
3483 @opindex Wstrict-overflow
3484 @opindex Wno-strict-overflow
3485 This option is only active when @option{-fstrict-overflow} is active.
3486 It warns about cases where the compiler optimizes based on the
3487 assumption that signed overflow does not occur. Note that it does not
3488 warn about all cases where the code might overflow: it only warns
3489 about cases where the compiler implements some optimization. Thus
3490 this warning depends on the optimization level.
3492 An optimization which assumes that signed overflow does not occur is
3493 perfectly safe if the values of the variables involved are such that
3494 overflow never does, in fact, occur. Therefore this warning can
3495 easily give a false positive: a warning about code which is not
3496 actually a problem. To help focus on important issues, several
3497 warning levels are defined. No warnings are issued for the use of
3498 undefined signed overflow when estimating how many iterations a loop
3499 will require, in particular when determining whether a loop will be
3503 @item -Wstrict-overflow=1
3504 Warn about cases which are both questionable and easy to avoid. For
3505 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3506 compiler will simplify this to @code{1}. This level of
3507 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3508 are not, and must be explicitly requested.
3510 @item -Wstrict-overflow=2
3511 Also warn about other cases where a comparison is simplified to a
3512 constant. For example: @code{abs (x) >= 0}. This can only be
3513 simplified when @option{-fstrict-overflow} is in effect, because
3514 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3515 zero. @option{-Wstrict-overflow} (with no level) is the same as
3516 @option{-Wstrict-overflow=2}.
3518 @item -Wstrict-overflow=3
3519 Also warn about other cases where a comparison is simplified. For
3520 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3522 @item -Wstrict-overflow=4
3523 Also warn about other simplifications not covered by the above cases.
3524 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3526 @item -Wstrict-overflow=5
3527 Also warn about cases where the compiler reduces the magnitude of a
3528 constant involved in a comparison. For example: @code{x + 2 > y} will
3529 be simplified to @code{x + 1 >= y}. This is reported only at the
3530 highest warning level because this simplification applies to many
3531 comparisons, so this warning level will give a very large number of
3535 @item -Warray-bounds
3536 @opindex Wno-array-bounds
3537 @opindex Warray-bounds
3538 This option is only active when @option{-ftree-vrp} is active
3539 (default for -O2 and above). It warns about subscripts to arrays
3540 that are always out of bounds. This warning is enabled by @option{-Wall}.
3542 @item -Wno-div-by-zero
3543 @opindex Wno-div-by-zero
3544 @opindex Wdiv-by-zero
3545 Do not warn about compile-time integer division by zero. Floating point
3546 division by zero is not warned about, as it can be a legitimate way of
3547 obtaining infinities and NaNs.
3549 @item -Wsystem-headers
3550 @opindex Wsystem-headers
3551 @opindex Wno-system-headers
3552 @cindex warnings from system headers
3553 @cindex system headers, warnings from
3554 Print warning messages for constructs found in system header files.
3555 Warnings from system headers are normally suppressed, on the assumption
3556 that they usually do not indicate real problems and would only make the
3557 compiler output harder to read. Using this command line option tells
3558 GCC to emit warnings from system headers as if they occurred in user
3559 code. However, note that using @option{-Wall} in conjunction with this
3560 option will @emph{not} warn about unknown pragmas in system
3561 headers---for that, @option{-Wunknown-pragmas} must also be used.
3564 @opindex Wfloat-equal
3565 @opindex Wno-float-equal
3566 Warn if floating point values are used in equality comparisons.
3568 The idea behind this is that sometimes it is convenient (for the
3569 programmer) to consider floating-point values as approximations to
3570 infinitely precise real numbers. If you are doing this, then you need
3571 to compute (by analyzing the code, or in some other way) the maximum or
3572 likely maximum error that the computation introduces, and allow for it
3573 when performing comparisons (and when producing output, but that's a
3574 different problem). In particular, instead of testing for equality, you
3575 would check to see whether the two values have ranges that overlap; and
3576 this is done with the relational operators, so equality comparisons are
3579 @item -Wtraditional @r{(C and Objective-C only)}
3580 @opindex Wtraditional
3581 @opindex Wno-traditional
3582 Warn about certain constructs that behave differently in traditional and
3583 ISO C@. Also warn about ISO C constructs that have no traditional C
3584 equivalent, and/or problematic constructs which should be avoided.
3588 Macro parameters that appear within string literals in the macro body.
3589 In traditional C macro replacement takes place within string literals,
3590 but does not in ISO C@.
3593 In traditional C, some preprocessor directives did not exist.
3594 Traditional preprocessors would only consider a line to be a directive
3595 if the @samp{#} appeared in column 1 on the line. Therefore
3596 @option{-Wtraditional} warns about directives that traditional C
3597 understands but would ignore because the @samp{#} does not appear as the
3598 first character on the line. It also suggests you hide directives like
3599 @samp{#pragma} not understood by traditional C by indenting them. Some
3600 traditional implementations would not recognize @samp{#elif}, so it
3601 suggests avoiding it altogether.
3604 A function-like macro that appears without arguments.
3607 The unary plus operator.
3610 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3611 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3612 constants.) Note, these suffixes appear in macros defined in the system
3613 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3614 Use of these macros in user code might normally lead to spurious
3615 warnings, however GCC's integrated preprocessor has enough context to
3616 avoid warning in these cases.
3619 A function declared external in one block and then used after the end of
3623 A @code{switch} statement has an operand of type @code{long}.
3626 A non-@code{static} function declaration follows a @code{static} one.
3627 This construct is not accepted by some traditional C compilers.
3630 The ISO type of an integer constant has a different width or
3631 signedness from its traditional type. This warning is only issued if
3632 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3633 typically represent bit patterns, are not warned about.
3636 Usage of ISO string concatenation is detected.
3639 Initialization of automatic aggregates.
3642 Identifier conflicts with labels. Traditional C lacks a separate
3643 namespace for labels.
3646 Initialization of unions. If the initializer is zero, the warning is
3647 omitted. This is done under the assumption that the zero initializer in
3648 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3649 initializer warnings and relies on default initialization to zero in the
3653 Conversions by prototypes between fixed/floating point values and vice
3654 versa. The absence of these prototypes when compiling with traditional
3655 C would cause serious problems. This is a subset of the possible
3656 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3659 Use of ISO C style function definitions. This warning intentionally is
3660 @emph{not} issued for prototype declarations or variadic functions
3661 because these ISO C features will appear in your code when using
3662 libiberty's traditional C compatibility macros, @code{PARAMS} and
3663 @code{VPARAMS}. This warning is also bypassed for nested functions
3664 because that feature is already a GCC extension and thus not relevant to
3665 traditional C compatibility.
3668 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3669 @opindex Wtraditional-conversion
3670 @opindex Wno-traditional-conversion
3671 Warn if a prototype causes a type conversion that is different from what
3672 would happen to the same argument in the absence of a prototype. This
3673 includes conversions of fixed point to floating and vice versa, and
3674 conversions changing the width or signedness of a fixed point argument
3675 except when the same as the default promotion.
3677 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3678 @opindex Wdeclaration-after-statement
3679 @opindex Wno-declaration-after-statement
3680 Warn when a declaration is found after a statement in a block. This
3681 construct, known from C++, was introduced with ISO C99 and is by default
3682 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3683 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3688 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3690 @item -Wno-endif-labels
3691 @opindex Wno-endif-labels
3692 @opindex Wendif-labels
3693 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3698 Warn whenever a local variable shadows another local variable, parameter or
3699 global variable or whenever a built-in function is shadowed.
3701 @item -Wlarger-than=@var{len}
3702 @opindex Wlarger-than=@var{len}
3703 @opindex Wlarger-than-@var{len}
3704 Warn whenever an object of larger than @var{len} bytes is defined.
3706 @item -Wframe-larger-than=@var{len}
3707 @opindex Wframe-larger-than
3708 Warn if the size of a function frame is larger than @var{len} bytes.
3709 The computation done to determine the stack frame size is approximate
3710 and not conservative.
3711 The actual requirements may be somewhat greater than @var{len}
3712 even if you do not get a warning. In addition, any space allocated
3713 via @code{alloca}, variable-length arrays, or related constructs
3714 is not included by the compiler when determining
3715 whether or not to issue a warning.
3717 @item -Wunsafe-loop-optimizations
3718 @opindex Wunsafe-loop-optimizations
3719 @opindex Wno-unsafe-loop-optimizations
3720 Warn if the loop cannot be optimized because the compiler could not
3721 assume anything on the bounds of the loop indices. With
3722 @option{-funsafe-loop-optimizations} warn if the compiler made
3725 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
3726 @opindex Wno-pedantic-ms-format
3727 @opindex Wpedantic-ms-format
3728 Disables the warnings about non-ISO @code{printf} / @code{scanf} format
3729 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets
3730 depending on the MS runtime, when you are using the options @option{-Wformat}
3731 and @option{-pedantic} without gnu-extensions.
3733 @item -Wpointer-arith
3734 @opindex Wpointer-arith
3735 @opindex Wno-pointer-arith
3736 Warn about anything that depends on the ``size of'' a function type or
3737 of @code{void}. GNU C assigns these types a size of 1, for
3738 convenience in calculations with @code{void *} pointers and pointers
3739 to functions. In C++, warn also when an arithmetic operation involves
3740 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3743 @opindex Wtype-limits
3744 @opindex Wno-type-limits
3745 Warn if a comparison is always true or always false due to the limited
3746 range of the data type, but do not warn for constant expressions. For
3747 example, warn if an unsigned variable is compared against zero with
3748 @samp{<} or @samp{>=}. This warning is also enabled by
3751 @item -Wbad-function-cast @r{(C and Objective-C only)}
3752 @opindex Wbad-function-cast
3753 @opindex Wno-bad-function-cast
3754 Warn whenever a function call is cast to a non-matching type.
3755 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3757 @item -Wc++-compat @r{(C and Objective-C only)}
3758 Warn about ISO C constructs that are outside of the common subset of
3759 ISO C and ISO C++, e.g.@: request for implicit conversion from
3760 @code{void *} to a pointer to non-@code{void} type.
3762 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3763 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3764 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3765 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3769 @opindex Wno-cast-qual
3770 Warn whenever a pointer is cast so as to remove a type qualifier from
3771 the target type. For example, warn if a @code{const char *} is cast
3772 to an ordinary @code{char *}.
3774 Also warn when making a cast which introduces a type qualifier in an
3775 unsafe way. For example, casting @code{char **} to @code{const char **}
3776 is unsafe, as in this example:
3779 /* p is char ** value. */
3780 const char **q = (const char **) p;
3781 /* Assignment of readonly string to const char * is OK. */
3783 /* Now char** pointer points to read-only memory. */
3788 @opindex Wcast-align
3789 @opindex Wno-cast-align
3790 Warn whenever a pointer is cast such that the required alignment of the
3791 target is increased. For example, warn if a @code{char *} is cast to
3792 an @code{int *} on machines where integers can only be accessed at
3793 two- or four-byte boundaries.
3795 @item -Wwrite-strings
3796 @opindex Wwrite-strings
3797 @opindex Wno-write-strings
3798 When compiling C, give string constants the type @code{const
3799 char[@var{length}]} so that copying the address of one into a
3800 non-@code{const} @code{char *} pointer will get a warning. These
3801 warnings will help you find at compile time code that can try to write
3802 into a string constant, but only if you have been very careful about
3803 using @code{const} in declarations and prototypes. Otherwise, it will
3804 just be a nuisance. This is why we did not make @option{-Wall} request
3807 When compiling C++, warn about the deprecated conversion from string
3808 literals to @code{char *}. This warning is enabled by default for C++
3813 @opindex Wno-clobbered
3814 Warn for variables that might be changed by @samp{longjmp} or
3815 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3818 @opindex Wconversion
3819 @opindex Wno-conversion
3820 Warn for implicit conversions that may alter a value. This includes
3821 conversions between real and integer, like @code{abs (x)} when
3822 @code{x} is @code{double}; conversions between signed and unsigned,
3823 like @code{unsigned ui = -1}; and conversions to smaller types, like
3824 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3825 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3826 changed by the conversion like in @code{abs (2.0)}. Warnings about
3827 conversions between signed and unsigned integers can be disabled by
3828 using @option{-Wno-sign-conversion}.
3830 For C++, also warn for conversions between @code{NULL} and non-pointer
3831 types; confusing overload resolution for user-defined conversions; and
3832 conversions that will never use a type conversion operator:
3833 conversions to @code{void}, the same type, a base class or a reference
3834 to them. Warnings about conversions between signed and unsigned
3835 integers are disabled by default in C++ unless
3836 @option{-Wsign-conversion} is explicitly enabled.
3839 @opindex Wempty-body
3840 @opindex Wno-empty-body
3841 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3842 while} statement. This warning is also enabled by @option{-Wextra}.
3844 @item -Wenum-compare
3845 @opindex Wenum-compare
3846 @opindex Wno-enum-compare
3847 Warn about a comparison between values of different enum types. In C++
3848 this warning is enabled by default. In C this warning is enabled by
3851 @item -Wjump-misses-init @r{(C, Objective-C only)}
3852 @opindex Wjump-misses-init
3853 @opindex Wno-jump-misses-init
3854 Warn if a @code{goto} statement or a @code{switch} statement jumps
3855 forward across the initialization of a variable, or jumps backward to a
3856 label after the variable has been initialized. This only warns about
3857 variables which are initialized when they are declared. This warning is
3858 only supported for C and Objective C; in C++ this sort of branch is an
3861 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
3862 can be disabled with the @option{-Wno-jump-misses-init} option.
3864 @item -Wsign-compare
3865 @opindex Wsign-compare
3866 @opindex Wno-sign-compare
3867 @cindex warning for comparison of signed and unsigned values
3868 @cindex comparison of signed and unsigned values, warning
3869 @cindex signed and unsigned values, comparison warning
3870 Warn when a comparison between signed and unsigned values could produce
3871 an incorrect result when the signed value is converted to unsigned.
3872 This warning is also enabled by @option{-Wextra}; to get the other warnings
3873 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3875 @item -Wsign-conversion
3876 @opindex Wsign-conversion
3877 @opindex Wno-sign-conversion
3878 Warn for implicit conversions that may change the sign of an integer
3879 value, like assigning a signed integer expression to an unsigned
3880 integer variable. An explicit cast silences the warning. In C, this
3881 option is enabled also by @option{-Wconversion}.
3885 @opindex Wno-address
3886 Warn about suspicious uses of memory addresses. These include using
3887 the address of a function in a conditional expression, such as
3888 @code{void func(void); if (func)}, and comparisons against the memory
3889 address of a string literal, such as @code{if (x == "abc")}. Such
3890 uses typically indicate a programmer error: the address of a function
3891 always evaluates to true, so their use in a conditional usually
3892 indicate that the programmer forgot the parentheses in a function
3893 call; and comparisons against string literals result in unspecified
3894 behavior and are not portable in C, so they usually indicate that the
3895 programmer intended to use @code{strcmp}. This warning is enabled by
3899 @opindex Wlogical-op
3900 @opindex Wno-logical-op
3901 Warn about suspicious uses of logical operators in expressions.
3902 This includes using logical operators in contexts where a
3903 bit-wise operator is likely to be expected.
3905 @item -Waggregate-return
3906 @opindex Waggregate-return
3907 @opindex Wno-aggregate-return
3908 Warn if any functions that return structures or unions are defined or
3909 called. (In languages where you can return an array, this also elicits
3912 @item -Wno-attributes
3913 @opindex Wno-attributes
3914 @opindex Wattributes
3915 Do not warn if an unexpected @code{__attribute__} is used, such as
3916 unrecognized attributes, function attributes applied to variables,
3917 etc. This will not stop errors for incorrect use of supported
3920 @item -Wno-builtin-macro-redefined
3921 @opindex Wno-builtin-macro-redefined
3922 @opindex Wbuiltin-macro-redefined
3923 Do not warn if certain built-in macros are redefined. This suppresses
3924 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
3925 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
3927 @item -Wstrict-prototypes @r{(C and Objective-C only)}
3928 @opindex Wstrict-prototypes
3929 @opindex Wno-strict-prototypes
3930 Warn if a function is declared or defined without specifying the
3931 argument types. (An old-style function definition is permitted without
3932 a warning if preceded by a declaration which specifies the argument
3935 @item -Wold-style-declaration @r{(C and Objective-C only)}
3936 @opindex Wold-style-declaration
3937 @opindex Wno-old-style-declaration
3938 Warn for obsolescent usages, according to the C Standard, in a
3939 declaration. For example, warn if storage-class specifiers like
3940 @code{static} are not the first things in a declaration. This warning
3941 is also enabled by @option{-Wextra}.
3943 @item -Wold-style-definition @r{(C and Objective-C only)}
3944 @opindex Wold-style-definition
3945 @opindex Wno-old-style-definition
3946 Warn if an old-style function definition is used. A warning is given
3947 even if there is a previous prototype.
3949 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
3950 @opindex Wmissing-parameter-type
3951 @opindex Wno-missing-parameter-type
3952 A function parameter is declared without a type specifier in K&R-style
3959 This warning is also enabled by @option{-Wextra}.
3961 @item -Wmissing-prototypes @r{(C and Objective-C only)}
3962 @opindex Wmissing-prototypes
3963 @opindex Wno-missing-prototypes
3964 Warn if a global function is defined without a previous prototype
3965 declaration. This warning is issued even if the definition itself
3966 provides a prototype. The aim is to detect global functions that fail
3967 to be declared in header files.
3969 @item -Wmissing-declarations
3970 @opindex Wmissing-declarations
3971 @opindex Wno-missing-declarations
3972 Warn if a global function is defined without a previous declaration.
3973 Do so even if the definition itself provides a prototype.
3974 Use this option to detect global functions that are not declared in
3975 header files. In C++, no warnings are issued for function templates,
3976 or for inline functions, or for functions in anonymous namespaces.
3978 @item -Wmissing-field-initializers
3979 @opindex Wmissing-field-initializers
3980 @opindex Wno-missing-field-initializers
3984 Warn if a structure's initializer has some fields missing. For
3985 example, the following code would cause such a warning, because
3986 @code{x.h} is implicitly zero:
3989 struct s @{ int f, g, h; @};
3990 struct s x = @{ 3, 4 @};
3993 This option does not warn about designated initializers, so the following
3994 modification would not trigger a warning:
3997 struct s @{ int f, g, h; @};
3998 struct s x = @{ .f = 3, .g = 4 @};
4001 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
4002 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
4004 @item -Wmissing-noreturn
4005 @opindex Wmissing-noreturn
4006 @opindex Wno-missing-noreturn
4007 Warn about functions which might be candidates for attribute @code{noreturn}.
4008 Note these are only possible candidates, not absolute ones. Care should
4009 be taken to manually verify functions actually do not ever return before
4010 adding the @code{noreturn} attribute, otherwise subtle code generation
4011 bugs could be introduced. You will not get a warning for @code{main} in
4012 hosted C environments.
4014 @item -Wmissing-format-attribute
4015 @opindex Wmissing-format-attribute
4016 @opindex Wno-missing-format-attribute
4019 Warn about function pointers which might be candidates for @code{format}
4020 attributes. Note these are only possible candidates, not absolute ones.
4021 GCC will guess that function pointers with @code{format} attributes that
4022 are used in assignment, initialization, parameter passing or return
4023 statements should have a corresponding @code{format} attribute in the
4024 resulting type. I.e.@: the left-hand side of the assignment or
4025 initialization, the type of the parameter variable, or the return type
4026 of the containing function respectively should also have a @code{format}
4027 attribute to avoid the warning.
4029 GCC will also warn about function definitions which might be
4030 candidates for @code{format} attributes. Again, these are only
4031 possible candidates. GCC will guess that @code{format} attributes
4032 might be appropriate for any function that calls a function like
4033 @code{vprintf} or @code{vscanf}, but this might not always be the
4034 case, and some functions for which @code{format} attributes are
4035 appropriate may not be detected.
4037 @item -Wno-multichar
4038 @opindex Wno-multichar
4040 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
4041 Usually they indicate a typo in the user's code, as they have
4042 implementation-defined values, and should not be used in portable code.
4044 @item -Wnormalized=<none|id|nfc|nfkc>
4045 @opindex Wnormalized=
4048 @cindex character set, input normalization
4049 In ISO C and ISO C++, two identifiers are different if they are
4050 different sequences of characters. However, sometimes when characters
4051 outside the basic ASCII character set are used, you can have two
4052 different character sequences that look the same. To avoid confusion,
4053 the ISO 10646 standard sets out some @dfn{normalization rules} which
4054 when applied ensure that two sequences that look the same are turned into
4055 the same sequence. GCC can warn you if you are using identifiers which
4056 have not been normalized; this option controls that warning.
4058 There are four levels of warning that GCC supports. The default is
4059 @option{-Wnormalized=nfc}, which warns about any identifier which is
4060 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
4061 recommended form for most uses.
4063 Unfortunately, there are some characters which ISO C and ISO C++ allow
4064 in identifiers that when turned into NFC aren't allowable as
4065 identifiers. That is, there's no way to use these symbols in portable
4066 ISO C or C++ and have all your identifiers in NFC@.
4067 @option{-Wnormalized=id} suppresses the warning for these characters.
4068 It is hoped that future versions of the standards involved will correct
4069 this, which is why this option is not the default.
4071 You can switch the warning off for all characters by writing
4072 @option{-Wnormalized=none}. You would only want to do this if you
4073 were using some other normalization scheme (like ``D''), because
4074 otherwise you can easily create bugs that are literally impossible to see.
4076 Some characters in ISO 10646 have distinct meanings but look identical
4077 in some fonts or display methodologies, especially once formatting has
4078 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
4079 LETTER N'', will display just like a regular @code{n} which has been
4080 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
4081 normalization scheme to convert all these into a standard form as
4082 well, and GCC will warn if your code is not in NFKC if you use
4083 @option{-Wnormalized=nfkc}. This warning is comparable to warning
4084 about every identifier that contains the letter O because it might be
4085 confused with the digit 0, and so is not the default, but may be
4086 useful as a local coding convention if the programming environment is
4087 unable to be fixed to display these characters distinctly.
4089 @item -Wno-deprecated
4090 @opindex Wno-deprecated
4091 @opindex Wdeprecated
4092 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
4094 @item -Wno-deprecated-declarations
4095 @opindex Wno-deprecated-declarations
4096 @opindex Wdeprecated-declarations
4097 Do not warn about uses of functions (@pxref{Function Attributes}),
4098 variables (@pxref{Variable Attributes}), and types (@pxref{Type
4099 Attributes}) marked as deprecated by using the @code{deprecated}
4103 @opindex Wno-overflow
4105 Do not warn about compile-time overflow in constant expressions.
4107 @item -Woverride-init @r{(C and Objective-C only)}
4108 @opindex Woverride-init
4109 @opindex Wno-override-init
4113 Warn if an initialized field without side effects is overridden when
4114 using designated initializers (@pxref{Designated Inits, , Designated
4117 This warning is included in @option{-Wextra}. To get other
4118 @option{-Wextra} warnings without this one, use @samp{-Wextra
4119 -Wno-override-init}.
4124 Warn if a structure is given the packed attribute, but the packed
4125 attribute has no effect on the layout or size of the structure.
4126 Such structures may be mis-aligned for little benefit. For
4127 instance, in this code, the variable @code{f.x} in @code{struct bar}
4128 will be misaligned even though @code{struct bar} does not itself
4129 have the packed attribute:
4136 @} __attribute__((packed));
4144 @item -Wpacked-bitfield-compat
4145 @opindex Wpacked-bitfield-compat
4146 @opindex Wno-packed-bitfield-compat
4147 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
4148 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
4149 the change can lead to differences in the structure layout. GCC
4150 informs you when the offset of such a field has changed in GCC 4.4.
4151 For example there is no longer a 4-bit padding between field @code{a}
4152 and @code{b} in this structure:
4159 @} __attribute__ ((packed));
4162 This warning is enabled by default. Use
4163 @option{-Wno-packed-bitfield-compat} to disable this warning.
4168 Warn if padding is included in a structure, either to align an element
4169 of the structure or to align the whole structure. Sometimes when this
4170 happens it is possible to rearrange the fields of the structure to
4171 reduce the padding and so make the structure smaller.
4173 @item -Wredundant-decls
4174 @opindex Wredundant-decls
4175 @opindex Wno-redundant-decls
4176 Warn if anything is declared more than once in the same scope, even in
4177 cases where multiple declaration is valid and changes nothing.
4179 @item -Wnested-externs @r{(C and Objective-C only)}
4180 @opindex Wnested-externs
4181 @opindex Wno-nested-externs
4182 Warn if an @code{extern} declaration is encountered within a function.
4184 @item -Wunreachable-code
4185 @opindex Wunreachable-code
4186 @opindex Wno-unreachable-code
4187 Warn if the compiler detects that code will never be executed.
4189 This option is intended to warn when the compiler detects that at
4190 least a whole line of source code will never be executed, because
4191 some condition is never satisfied or because it is after a
4192 procedure that never returns.
4194 It is possible for this option to produce a warning even though there
4195 are circumstances under which part of the affected line can be executed,
4196 so care should be taken when removing apparently-unreachable code.
4198 For instance, when a function is inlined, a warning may mean that the
4199 line is unreachable in only one inlined copy of the function.
4201 This option is not made part of @option{-Wall} because in a debugging
4202 version of a program there is often substantial code which checks
4203 correct functioning of the program and is, hopefully, unreachable
4204 because the program does work. Another common use of unreachable
4205 code is to provide behavior which is selectable at compile-time.
4210 Warn if a function can not be inlined and it was declared as inline.
4211 Even with this option, the compiler will not warn about failures to
4212 inline functions declared in system headers.
4214 The compiler uses a variety of heuristics to determine whether or not
4215 to inline a function. For example, the compiler takes into account
4216 the size of the function being inlined and the amount of inlining
4217 that has already been done in the current function. Therefore,
4218 seemingly insignificant changes in the source program can cause the
4219 warnings produced by @option{-Winline} to appear or disappear.
4221 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
4222 @opindex Wno-invalid-offsetof
4223 @opindex Winvalid-offsetof
4224 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
4225 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4226 to a non-POD type is undefined. In existing C++ implementations,
4227 however, @samp{offsetof} typically gives meaningful results even when
4228 applied to certain kinds of non-POD types. (Such as a simple
4229 @samp{struct} that fails to be a POD type only by virtue of having a
4230 constructor.) This flag is for users who are aware that they are
4231 writing nonportable code and who have deliberately chosen to ignore the
4234 The restrictions on @samp{offsetof} may be relaxed in a future version
4235 of the C++ standard.
4237 @item -Wno-int-to-pointer-cast @r{(C and Objective-C only)}
4238 @opindex Wno-int-to-pointer-cast
4239 @opindex Wint-to-pointer-cast
4240 Suppress warnings from casts to pointer type of an integer of a
4243 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4244 @opindex Wno-pointer-to-int-cast
4245 @opindex Wpointer-to-int-cast
4246 Suppress warnings from casts from a pointer to an integer type of a
4250 @opindex Winvalid-pch
4251 @opindex Wno-invalid-pch
4252 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4253 the search path but can't be used.
4257 @opindex Wno-long-long
4258 Warn if @samp{long long} type is used. This is enabled by either
4259 @option{-pedantic} or @option{-Wtraditional} in ISO C90 and C++98
4260 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
4262 @item -Wvariadic-macros
4263 @opindex Wvariadic-macros
4264 @opindex Wno-variadic-macros
4265 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4266 alternate syntax when in pedantic ISO C99 mode. This is default.
4267 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4272 Warn if variable length array is used in the code.
4273 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4274 the variable length array.
4276 @item -Wvolatile-register-var
4277 @opindex Wvolatile-register-var
4278 @opindex Wno-volatile-register-var
4279 Warn if a register variable is declared volatile. The volatile
4280 modifier does not inhibit all optimizations that may eliminate reads
4281 and/or writes to register variables. This warning is enabled by
4284 @item -Wdisabled-optimization
4285 @opindex Wdisabled-optimization
4286 @opindex Wno-disabled-optimization
4287 Warn if a requested optimization pass is disabled. This warning does
4288 not generally indicate that there is anything wrong with your code; it
4289 merely indicates that GCC's optimizers were unable to handle the code
4290 effectively. Often, the problem is that your code is too big or too
4291 complex; GCC will refuse to optimize programs when the optimization
4292 itself is likely to take inordinate amounts of time.
4294 @item -Wpointer-sign @r{(C and Objective-C only)}
4295 @opindex Wpointer-sign
4296 @opindex Wno-pointer-sign
4297 Warn for pointer argument passing or assignment with different signedness.
4298 This option is only supported for C and Objective-C@. It is implied by
4299 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4300 @option{-Wno-pointer-sign}.
4302 @item -Wstack-protector
4303 @opindex Wstack-protector
4304 @opindex Wno-stack-protector
4305 This option is only active when @option{-fstack-protector} is active. It
4306 warns about functions that will not be protected against stack smashing.
4309 @opindex Wno-mudflap
4310 Suppress warnings about constructs that cannot be instrumented by
4313 @item -Woverlength-strings
4314 @opindex Woverlength-strings
4315 @opindex Wno-overlength-strings
4316 Warn about string constants which are longer than the ``minimum
4317 maximum'' length specified in the C standard. Modern compilers
4318 generally allow string constants which are much longer than the
4319 standard's minimum limit, but very portable programs should avoid
4320 using longer strings.
4322 The limit applies @emph{after} string constant concatenation, and does
4323 not count the trailing NUL@. In C89, the limit was 509 characters; in
4324 C99, it was raised to 4095. C++98 does not specify a normative
4325 minimum maximum, so we do not diagnose overlength strings in C++@.
4327 This option is implied by @option{-pedantic}, and can be disabled with
4328 @option{-Wno-overlength-strings}.
4330 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
4331 @opindex Wunsuffixed-float-constants
4333 GCC will issue a warning for any floating constant that does not have
4334 a suffix. When used together with @option{-Wsystem-headers} it will
4335 warn about such constants in system header files. This can be useful
4336 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
4337 from the decimal floating-point extension to C99.
4340 @node Debugging Options
4341 @section Options for Debugging Your Program or GCC
4342 @cindex options, debugging
4343 @cindex debugging information options
4345 GCC has various special options that are used for debugging
4346 either your program or GCC:
4351 Produce debugging information in the operating system's native format
4352 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4355 On most systems that use stabs format, @option{-g} enables use of extra
4356 debugging information that only GDB can use; this extra information
4357 makes debugging work better in GDB but will probably make other debuggers
4359 refuse to read the program. If you want to control for certain whether
4360 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4361 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4363 GCC allows you to use @option{-g} with
4364 @option{-O}. The shortcuts taken by optimized code may occasionally
4365 produce surprising results: some variables you declared may not exist
4366 at all; flow of control may briefly move where you did not expect it;
4367 some statements may not be executed because they compute constant
4368 results or their values were already at hand; some statements may
4369 execute in different places because they were moved out of loops.
4371 Nevertheless it proves possible to debug optimized output. This makes
4372 it reasonable to use the optimizer for programs that might have bugs.
4374 The following options are useful when GCC is generated with the
4375 capability for more than one debugging format.
4379 Produce debugging information for use by GDB@. This means to use the
4380 most expressive format available (DWARF 2, stabs, or the native format
4381 if neither of those are supported), including GDB extensions if at all
4386 Produce debugging information in stabs format (if that is supported),
4387 without GDB extensions. This is the format used by DBX on most BSD
4388 systems. On MIPS, Alpha and System V Release 4 systems this option
4389 produces stabs debugging output which is not understood by DBX or SDB@.
4390 On System V Release 4 systems this option requires the GNU assembler.
4392 @item -feliminate-unused-debug-symbols
4393 @opindex feliminate-unused-debug-symbols
4394 Produce debugging information in stabs format (if that is supported),
4395 for only symbols that are actually used.
4397 @item -femit-class-debug-always
4398 Instead of emitting debugging information for a C++ class in only one
4399 object file, emit it in all object files using the class. This option
4400 should be used only with debuggers that are unable to handle the way GCC
4401 normally emits debugging information for classes because using this
4402 option will increase the size of debugging information by as much as a
4407 Produce debugging information in stabs format (if that is supported),
4408 using GNU extensions understood only by the GNU debugger (GDB)@. The
4409 use of these extensions is likely to make other debuggers crash or
4410 refuse to read the program.
4414 Produce debugging information in COFF format (if that is supported).
4415 This is the format used by SDB on most System V systems prior to
4420 Produce debugging information in XCOFF format (if that is supported).
4421 This is the format used by the DBX debugger on IBM RS/6000 systems.
4425 Produce debugging information in XCOFF format (if that is supported),
4426 using GNU extensions understood only by the GNU debugger (GDB)@. The
4427 use of these extensions is likely to make other debuggers crash or
4428 refuse to read the program, and may cause assemblers other than the GNU
4429 assembler (GAS) to fail with an error.
4431 @item -gdwarf-@var{version}
4432 @opindex gdwarf-@var{version}
4433 Produce debugging information in DWARF format (if that is
4434 supported). This is the format used by DBX on IRIX 6. The value
4435 of @var{version} may be either 2, 3 or 4; the default version is 2.
4437 Note that with DWARF version 2 some ports require, and will always
4438 use, some non-conflicting DWARF 3 extensions in the unwind tables.
4440 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
4441 for maximum benefit.
4443 @item -gstrict-dwarf
4444 @opindex gstrict-dwarf
4445 Disallow using extensions of later DWARF standard version than selected
4446 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
4447 DWARF extensions from later standard versions is allowed.
4449 @item -gno-strict-dwarf
4450 @opindex gno-strict-dwarf
4451 Allow using extensions of later DWARF standard version than selected with
4452 @option{-gdwarf-@var{version}}.
4456 Produce debugging information in VMS debug format (if that is
4457 supported). This is the format used by DEBUG on VMS systems.
4460 @itemx -ggdb@var{level}
4461 @itemx -gstabs@var{level}
4462 @itemx -gcoff@var{level}
4463 @itemx -gxcoff@var{level}
4464 @itemx -gvms@var{level}
4465 Request debugging information and also use @var{level} to specify how
4466 much information. The default level is 2.
4468 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4471 Level 1 produces minimal information, enough for making backtraces in
4472 parts of the program that you don't plan to debug. This includes
4473 descriptions of functions and external variables, but no information
4474 about local variables and no line numbers.
4476 Level 3 includes extra information, such as all the macro definitions
4477 present in the program. Some debuggers support macro expansion when
4478 you use @option{-g3}.
4480 @option{-gdwarf-2} does not accept a concatenated debug level, because
4481 GCC used to support an option @option{-gdwarf} that meant to generate
4482 debug information in version 1 of the DWARF format (which is very
4483 different from version 2), and it would have been too confusing. That
4484 debug format is long obsolete, but the option cannot be changed now.
4485 Instead use an additional @option{-g@var{level}} option to change the
4486 debug level for DWARF.
4490 Turn off generation of debug info, if leaving out this option would have
4491 generated it, or turn it on at level 2 otherwise. The position of this
4492 argument in the command line does not matter, it takes effect after all
4493 other options are processed, and it does so only once, no matter how
4494 many times it is given. This is mainly intended to be used with
4495 @option{-fcompare-debug}.
4497 @item -fdump-final-insns@r{[}=@var{file}@r{]}
4498 @opindex fdump-final-insns
4499 Dump the final internal representation (RTL) to @var{file}. If the
4500 optional argument is omitted (or if @var{file} is @code{.}), the name
4501 of the dump file will be determined by appending @code{.gkd} to the
4502 compilation output file name.
4504 @item -fcompare-debug@r{[}=@var{opts}@r{]}
4505 @opindex fcompare-debug
4506 @opindex fno-compare-debug
4507 If no error occurs during compilation, run the compiler a second time,
4508 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
4509 passed to the second compilation. Dump the final internal
4510 representation in both compilations, and print an error if they differ.
4512 If the equal sign is omitted, the default @option{-gtoggle} is used.
4514 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
4515 and nonzero, implicitly enables @option{-fcompare-debug}. If
4516 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
4517 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
4520 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
4521 is equivalent to @option{-fno-compare-debug}, which disables the dumping
4522 of the final representation and the second compilation, preventing even
4523 @env{GCC_COMPARE_DEBUG} from taking effect.
4525 To verify full coverage during @option{-fcompare-debug} testing, set
4526 @env{GCC_COMPARE_DEBUG} to say @samp{-fcompare-debug-not-overridden},
4527 which GCC will reject as an invalid option in any actual compilation
4528 (rather than preprocessing, assembly or linking). To get just a
4529 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
4530 not overridden} will do.
4532 @item -fcompare-debug-second
4533 @opindex fcompare-debug-second
4534 This option is implicitly passed to the compiler for the second
4535 compilation requested by @option{-fcompare-debug}, along with options to
4536 silence warnings, and omitting other options that would cause
4537 side-effect compiler outputs to files or to the standard output. Dump
4538 files and preserved temporary files are renamed so as to contain the
4539 @code{.gk} additional extension during the second compilation, to avoid
4540 overwriting those generated by the first.
4542 When this option is passed to the compiler driver, it causes the
4543 @emph{first} compilation to be skipped, which makes it useful for little
4544 other than debugging the compiler proper.
4546 @item -feliminate-dwarf2-dups
4547 @opindex feliminate-dwarf2-dups
4548 Compress DWARF2 debugging information by eliminating duplicated
4549 information about each symbol. This option only makes sense when
4550 generating DWARF2 debugging information with @option{-gdwarf-2}.
4552 @item -femit-struct-debug-baseonly
4553 Emit debug information for struct-like types
4554 only when the base name of the compilation source file
4555 matches the base name of file in which the struct was defined.
4557 This option substantially reduces the size of debugging information,
4558 but at significant potential loss in type information to the debugger.
4559 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4560 See @option{-femit-struct-debug-detailed} for more detailed control.
4562 This option works only with DWARF 2.
4564 @item -femit-struct-debug-reduced
4565 Emit debug information for struct-like types
4566 only when the base name of the compilation source file
4567 matches the base name of file in which the type was defined,
4568 unless the struct is a template or defined in a system header.
4570 This option significantly reduces the size of debugging information,
4571 with some potential loss in type information to the debugger.
4572 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4573 See @option{-femit-struct-debug-detailed} for more detailed control.
4575 This option works only with DWARF 2.
4577 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4578 Specify the struct-like types
4579 for which the compiler will generate debug information.
4580 The intent is to reduce duplicate struct debug information
4581 between different object files within the same program.
4583 This option is a detailed version of
4584 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4585 which will serve for most needs.
4587 A specification has the syntax
4588 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4590 The optional first word limits the specification to
4591 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4592 A struct type is used directly when it is the type of a variable, member.
4593 Indirect uses arise through pointers to structs.
4594 That is, when use of an incomplete struct would be legal, the use is indirect.
4596 @samp{struct one direct; struct two * indirect;}.
4598 The optional second word limits the specification to
4599 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4600 Generic structs are a bit complicated to explain.
4601 For C++, these are non-explicit specializations of template classes,
4602 or non-template classes within the above.
4603 Other programming languages have generics,
4604 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4606 The third word specifies the source files for those
4607 structs for which the compiler will emit debug information.
4608 The values @samp{none} and @samp{any} have the normal meaning.
4609 The value @samp{base} means that
4610 the base of name of the file in which the type declaration appears
4611 must match the base of the name of the main compilation file.
4612 In practice, this means that
4613 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4614 but types declared in other header will not.
4615 The value @samp{sys} means those types satisfying @samp{base}
4616 or declared in system or compiler headers.
4618 You may need to experiment to determine the best settings for your application.
4620 The default is @samp{-femit-struct-debug-detailed=all}.
4622 This option works only with DWARF 2.
4624 @item -fenable-icf-debug
4625 @opindex fenable-icf-debug
4626 Generate additional debug information to support identical code folding (ICF).
4627 This option only works with DWARF version 2 or higher.
4629 @item -fno-merge-debug-strings
4630 @opindex fmerge-debug-strings
4631 @opindex fno-merge-debug-strings
4632 Direct the linker to not merge together strings in the debugging
4633 information which are identical in different object files. Merging is
4634 not supported by all assemblers or linkers. Merging decreases the size
4635 of the debug information in the output file at the cost of increasing
4636 link processing time. Merging is enabled by default.
4638 @item -fdebug-prefix-map=@var{old}=@var{new}
4639 @opindex fdebug-prefix-map
4640 When compiling files in directory @file{@var{old}}, record debugging
4641 information describing them as in @file{@var{new}} instead.
4643 @item -fno-dwarf2-cfi-asm
4644 @opindex fdwarf2-cfi-asm
4645 @opindex fno-dwarf2-cfi-asm
4646 Emit DWARF 2 unwind info as compiler generated @code{.eh_frame} section
4647 instead of using GAS @code{.cfi_*} directives.
4649 @cindex @command{prof}
4652 Generate extra code to write profile information suitable for the
4653 analysis program @command{prof}. You must use this option when compiling
4654 the source files you want data about, and you must also use it when
4657 @cindex @command{gprof}
4660 Generate extra code to write profile information suitable for the
4661 analysis program @command{gprof}. You must use this option when compiling
4662 the source files you want data about, and you must also use it when
4667 Makes the compiler print out each function name as it is compiled, and
4668 print some statistics about each pass when it finishes.
4671 @opindex ftime-report
4672 Makes the compiler print some statistics about the time consumed by each
4673 pass when it finishes.
4676 @opindex fmem-report
4677 Makes the compiler print some statistics about permanent memory
4678 allocation when it finishes.
4680 @item -fpre-ipa-mem-report
4681 @opindex fpre-ipa-mem-report
4682 @item -fpost-ipa-mem-report
4683 @opindex fpost-ipa-mem-report
4684 Makes the compiler print some statistics about permanent memory
4685 allocation before or after interprocedural optimization.
4687 @item -fprofile-arcs
4688 @opindex fprofile-arcs
4689 Add code so that program flow @dfn{arcs} are instrumented. During
4690 execution the program records how many times each branch and call is
4691 executed and how many times it is taken or returns. When the compiled
4692 program exits it saves this data to a file called
4693 @file{@var{auxname}.gcda} for each source file. The data may be used for
4694 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4695 test coverage analysis (@option{-ftest-coverage}). Each object file's
4696 @var{auxname} is generated from the name of the output file, if
4697 explicitly specified and it is not the final executable, otherwise it is
4698 the basename of the source file. In both cases any suffix is removed
4699 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4700 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4701 @xref{Cross-profiling}.
4703 @cindex @command{gcov}
4707 This option is used to compile and link code instrumented for coverage
4708 analysis. The option is a synonym for @option{-fprofile-arcs}
4709 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4710 linking). See the documentation for those options for more details.
4715 Compile the source files with @option{-fprofile-arcs} plus optimization
4716 and code generation options. For test coverage analysis, use the
4717 additional @option{-ftest-coverage} option. You do not need to profile
4718 every source file in a program.
4721 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4722 (the latter implies the former).
4725 Run the program on a representative workload to generate the arc profile
4726 information. This may be repeated any number of times. You can run
4727 concurrent instances of your program, and provided that the file system
4728 supports locking, the data files will be correctly updated. Also
4729 @code{fork} calls are detected and correctly handled (double counting
4733 For profile-directed optimizations, compile the source files again with
4734 the same optimization and code generation options plus
4735 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4736 Control Optimization}).
4739 For test coverage analysis, use @command{gcov} to produce human readable
4740 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4741 @command{gcov} documentation for further information.
4745 With @option{-fprofile-arcs}, for each function of your program GCC
4746 creates a program flow graph, then finds a spanning tree for the graph.
4747 Only arcs that are not on the spanning tree have to be instrumented: the
4748 compiler adds code to count the number of times that these arcs are
4749 executed. When an arc is the only exit or only entrance to a block, the
4750 instrumentation code can be added to the block; otherwise, a new basic
4751 block must be created to hold the instrumentation code.
4754 @item -ftest-coverage
4755 @opindex ftest-coverage
4756 Produce a notes file that the @command{gcov} code-coverage utility
4757 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4758 show program coverage. Each source file's note file is called
4759 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4760 above for a description of @var{auxname} and instructions on how to
4761 generate test coverage data. Coverage data will match the source files
4762 more closely, if you do not optimize.
4764 @item -fdbg-cnt-list
4765 @opindex fdbg-cnt-list
4766 Print the name and the counter upperbound for all debug counters.
4768 @item -fdbg-cnt=@var{counter-value-list}
4770 Set the internal debug counter upperbound. @var{counter-value-list}
4771 is a comma-separated list of @var{name}:@var{value} pairs
4772 which sets the upperbound of each debug counter @var{name} to @var{value}.
4773 All debug counters have the initial upperbound of @var{UINT_MAX},
4774 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4775 e.g. With -fdbg-cnt=dce:10,tail_call:0
4776 dbg_cnt(dce) will return true only for first 10 invocations
4777 and dbg_cnt(tail_call) will return false always.
4779 @item -d@var{letters}
4780 @itemx -fdump-rtl-@var{pass}
4782 Says to make debugging dumps during compilation at times specified by
4783 @var{letters}. This is used for debugging the RTL-based passes of the
4784 compiler. The file names for most of the dumps are made by appending
4785 a pass number and a word to the @var{dumpname}, and the files are
4786 created in the directory of the output file. @var{dumpname} is
4787 generated from the name of the output file, if explicitly specified
4788 and it is not an executable, otherwise it is the basename of the
4789 source file. These switches may have different effects when
4790 @option{-E} is used for preprocessing.
4792 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
4793 @option{-d} option @var{letters}. Here are the possible
4794 letters for use in @var{pass} and @var{letters}, and their meanings:
4798 @item -fdump-rtl-alignments
4799 @opindex fdump-rtl-alignments
4800 Dump after branch alignments have been computed.
4802 @item -fdump-rtl-asmcons
4803 @opindex fdump-rtl-asmcons
4804 Dump after fixing rtl statements that have unsatisfied in/out constraints.
4806 @item -fdump-rtl-auto_inc_dec
4807 @opindex fdump-rtl-auto_inc_dec
4808 Dump after auto-inc-dec discovery. This pass is only run on
4809 architectures that have auto inc or auto dec instructions.
4811 @item -fdump-rtl-barriers
4812 @opindex fdump-rtl-barriers
4813 Dump after cleaning up the barrier instructions.
4815 @item -fdump-rtl-bbpart
4816 @opindex fdump-rtl-bbpart
4817 Dump after partitioning hot and cold basic blocks.
4819 @item -fdump-rtl-bbro
4820 @opindex fdump-rtl-bbro
4821 Dump after block reordering.
4823 @item -fdump-rtl-btl1
4824 @itemx -fdump-rtl-btl2
4825 @opindex fdump-rtl-btl2
4826 @opindex fdump-rtl-btl2
4827 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
4828 after the two branch
4829 target load optimization passes.
4831 @item -fdump-rtl-bypass
4832 @opindex fdump-rtl-bypass
4833 Dump after jump bypassing and control flow optimizations.
4835 @item -fdump-rtl-combine
4836 @opindex fdump-rtl-combine
4837 Dump after the RTL instruction combination pass.
4839 @item -fdump-rtl-compgotos
4840 @opindex fdump-rtl-compgotos
4841 Dump after duplicating the computed gotos.
4843 @item -fdump-rtl-ce1
4844 @itemx -fdump-rtl-ce2
4845 @itemx -fdump-rtl-ce3
4846 @opindex fdump-rtl-ce1
4847 @opindex fdump-rtl-ce2
4848 @opindex fdump-rtl-ce3
4849 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
4850 @option{-fdump-rtl-ce3} enable dumping after the three
4851 if conversion passes.
4853 @itemx -fdump-rtl-cprop_hardreg
4854 @opindex fdump-rtl-cprop_hardreg
4855 Dump after hard register copy propagation.
4857 @itemx -fdump-rtl-csa
4858 @opindex fdump-rtl-csa
4859 Dump after combining stack adjustments.
4861 @item -fdump-rtl-cse1
4862 @itemx -fdump-rtl-cse2
4863 @opindex fdump-rtl-cse1
4864 @opindex fdump-rtl-cse2
4865 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
4866 the two common sub-expression elimination passes.
4868 @itemx -fdump-rtl-dce
4869 @opindex fdump-rtl-dce
4870 Dump after the standalone dead code elimination passes.
4872 @itemx -fdump-rtl-dbr
4873 @opindex fdump-rtl-dbr
4874 Dump after delayed branch scheduling.
4876 @item -fdump-rtl-dce1
4877 @itemx -fdump-rtl-dce2
4878 @opindex fdump-rtl-dce1
4879 @opindex fdump-rtl-dce2
4880 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
4881 the two dead store elimination passes.
4884 @opindex fdump-rtl-eh
4885 Dump after finalization of EH handling code.
4887 @item -fdump-rtl-eh_ranges
4888 @opindex fdump-rtl-eh_ranges
4889 Dump after conversion of EH handling range regions.
4891 @item -fdump-rtl-expand
4892 @opindex fdump-rtl-expand
4893 Dump after RTL generation.
4895 @item -fdump-rtl-fwprop1
4896 @itemx -fdump-rtl-fwprop2
4897 @opindex fdump-rtl-fwprop1
4898 @opindex fdump-rtl-fwprop2
4899 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
4900 dumping after the two forward propagation passes.
4902 @item -fdump-rtl-gcse1
4903 @itemx -fdump-rtl-gcse2
4904 @opindex fdump-rtl-gcse1
4905 @opindex fdump-rtl-gcse2
4906 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
4907 after global common subexpression elimination.
4909 @item -fdump-rtl-init-regs
4910 @opindex fdump-rtl-init-regs
4911 Dump after the initialization of the registers.
4913 @item -fdump-rtl-initvals
4914 @opindex fdump-rtl-initvals
4915 Dump after the computation of the initial value sets.
4917 @itemx -fdump-rtl-into_cfglayout
4918 @opindex fdump-rtl-into_cfglayout
4919 Dump after converting to cfglayout mode.
4921 @item -fdump-rtl-ira
4922 @opindex fdump-rtl-ira
4923 Dump after iterated register allocation.
4925 @item -fdump-rtl-jump
4926 @opindex fdump-rtl-jump
4927 Dump after the second jump optimization.
4929 @item -fdump-rtl-loop2
4930 @opindex fdump-rtl-loop2
4931 @option{-fdump-rtl-loop2} enables dumping after the rtl
4932 loop optimization passes.
4934 @item -fdump-rtl-mach
4935 @opindex fdump-rtl-mach
4936 Dump after performing the machine dependent reorganization pass, if that
4939 @item -fdump-rtl-mode_sw
4940 @opindex fdump-rtl-mode_sw
4941 Dump after removing redundant mode switches.
4943 @item -fdump-rtl-rnreg
4944 @opindex fdump-rtl-rnreg
4945 Dump after register renumbering.
4947 @itemx -fdump-rtl-outof_cfglayout
4948 @opindex fdump-rtl-outof_cfglayout
4949 Dump after converting from cfglayout mode.
4951 @item -fdump-rtl-peephole2
4952 @opindex fdump-rtl-peephole2
4953 Dump after the peephole pass.
4955 @item -fdump-rtl-postreload
4956 @opindex fdump-rtl-postreload
4957 Dump after post-reload optimizations.
4959 @itemx -fdump-rtl-pro_and_epilogue
4960 @opindex fdump-rtl-pro_and_epilogue
4961 Dump after generating the function pro and epilogues.
4963 @item -fdump-rtl-regmove
4964 @opindex fdump-rtl-regmove
4965 Dump after the register move pass.
4967 @item -fdump-rtl-sched1
4968 @itemx -fdump-rtl-sched2
4969 @opindex fdump-rtl-sched1
4970 @opindex fdump-rtl-sched2
4971 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
4972 after the basic block scheduling passes.
4974 @item -fdump-rtl-see
4975 @opindex fdump-rtl-see
4976 Dump after sign extension elimination.
4978 @item -fdump-rtl-seqabstr
4979 @opindex fdump-rtl-seqabstr
4980 Dump after common sequence discovery.
4982 @item -fdump-rtl-shorten
4983 @opindex fdump-rtl-shorten
4984 Dump after shortening branches.
4986 @item -fdump-rtl-sibling
4987 @opindex fdump-rtl-sibling
4988 Dump after sibling call optimizations.
4990 @item -fdump-rtl-split1
4991 @itemx -fdump-rtl-split2
4992 @itemx -fdump-rtl-split3
4993 @itemx -fdump-rtl-split4
4994 @itemx -fdump-rtl-split5
4995 @opindex fdump-rtl-split1
4996 @opindex fdump-rtl-split2
4997 @opindex fdump-rtl-split3
4998 @opindex fdump-rtl-split4
4999 @opindex fdump-rtl-split5
5000 @option{-fdump-rtl-split1}, @option{-fdump-rtl-split2},
5001 @option{-fdump-rtl-split3}, @option{-fdump-rtl-split4} and
5002 @option{-fdump-rtl-split5} enable dumping after five rounds of
5003 instruction splitting.
5005 @item -fdump-rtl-sms
5006 @opindex fdump-rtl-sms
5007 Dump after modulo scheduling. This pass is only run on some
5010 @item -fdump-rtl-stack
5011 @opindex fdump-rtl-stack
5012 Dump after conversion from GCC's "flat register file" registers to the
5013 x87's stack-like registers. This pass is only run on x86 variants.
5015 @item -fdump-rtl-subreg1
5016 @itemx -fdump-rtl-subreg2
5017 @opindex fdump-rtl-subreg1
5018 @opindex fdump-rtl-subreg2
5019 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
5020 the two subreg expansion passes.
5022 @item -fdump-rtl-unshare
5023 @opindex fdump-rtl-unshare
5024 Dump after all rtl has been unshared.
5026 @item -fdump-rtl-vartrack
5027 @opindex fdump-rtl-vartrack
5028 Dump after variable tracking.
5030 @item -fdump-rtl-vregs
5031 @opindex fdump-rtl-vregs
5032 Dump after converting virtual registers to hard registers.
5034 @item -fdump-rtl-web
5035 @opindex fdump-rtl-web
5036 Dump after live range splitting.
5038 @item -fdump-rtl-regclass
5039 @itemx -fdump-rtl-subregs_of_mode_init
5040 @itemx -fdump-rtl-subregs_of_mode_finish
5041 @itemx -fdump-rtl-dfinit
5042 @itemx -fdump-rtl-dfinish
5043 @opindex fdump-rtl-regclass
5044 @opindex fdump-rtl-subregs_of_mode_init
5045 @opindex fdump-rtl-subregs_of_mode_finish
5046 @opindex fdump-rtl-dfinit
5047 @opindex fdump-rtl-dfinish
5048 These dumps are defined but always produce empty files.
5050 @item -fdump-rtl-all
5051 @opindex fdump-rtl-all
5052 Produce all the dumps listed above.
5056 Annotate the assembler output with miscellaneous debugging information.
5060 Dump all macro definitions, at the end of preprocessing, in addition to
5065 Produce a core dump whenever an error occurs.
5069 Print statistics on memory usage, at the end of the run, to
5074 Annotate the assembler output with a comment indicating which
5075 pattern and alternative was used. The length of each instruction is
5080 Dump the RTL in the assembler output as a comment before each instruction.
5081 Also turns on @option{-dp} annotation.
5085 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
5086 dump a representation of the control flow graph suitable for viewing with VCG
5087 to @file{@var{file}.@var{pass}.vcg}.
5091 Just generate RTL for a function instead of compiling it. Usually used
5092 with @option{-fdump-rtl-expand}.
5096 Dump debugging information during parsing, to standard error.
5100 @opindex fdump-noaddr
5101 When doing debugging dumps, suppress address output. This makes it more
5102 feasible to use diff on debugging dumps for compiler invocations with
5103 different compiler binaries and/or different
5104 text / bss / data / heap / stack / dso start locations.
5106 @item -fdump-unnumbered
5107 @opindex fdump-unnumbered
5108 When doing debugging dumps, suppress instruction numbers and address output.
5109 This makes it more feasible to use diff on debugging dumps for compiler
5110 invocations with different options, in particular with and without
5113 @item -fdump-unnumbered-links
5114 @opindex fdump-unnumbered-links
5115 When doing debugging dumps (see @option{-d} option above), suppress
5116 instruction numbers for the links to the previous and next instructions
5119 @item -fdump-translation-unit @r{(C++ only)}
5120 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
5121 @opindex fdump-translation-unit
5122 Dump a representation of the tree structure for the entire translation
5123 unit to a file. The file name is made by appending @file{.tu} to the
5124 source file name, and the file is created in the same directory as the
5125 output file. If the @samp{-@var{options}} form is used, @var{options}
5126 controls the details of the dump as described for the
5127 @option{-fdump-tree} options.
5129 @item -fdump-class-hierarchy @r{(C++ only)}
5130 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
5131 @opindex fdump-class-hierarchy
5132 Dump a representation of each class's hierarchy and virtual function
5133 table layout to a file. The file name is made by appending
5134 @file{.class} to the source file name, and the file is created in the
5135 same directory as the output file. If the @samp{-@var{options}} form
5136 is used, @var{options} controls the details of the dump as described
5137 for the @option{-fdump-tree} options.
5139 @item -fdump-ipa-@var{switch}
5141 Control the dumping at various stages of inter-procedural analysis
5142 language tree to a file. The file name is generated by appending a
5143 switch specific suffix to the source file name, and the file is created
5144 in the same directory as the output file. The following dumps are
5149 Enables all inter-procedural analysis dumps.
5152 Dumps information about call-graph optimization, unused function removal,
5153 and inlining decisions.
5156 Dump after function inlining.
5160 @item -fdump-statistics-@var{option}
5161 @opindex fdump-statistics
5162 Enable and control dumping of pass statistics in a separate file. The
5163 file name is generated by appending a suffix ending in
5164 @samp{.statistics} to the source file name, and the file is created in
5165 the same directory as the output file. If the @samp{-@var{option}}
5166 form is used, @samp{-stats} will cause counters to be summed over the
5167 whole compilation unit while @samp{-details} will dump every event as
5168 the passes generate them. The default with no option is to sum
5169 counters for each function compiled.
5171 @item -fdump-tree-@var{switch}
5172 @itemx -fdump-tree-@var{switch}-@var{options}
5174 Control the dumping at various stages of processing the intermediate
5175 language tree to a file. The file name is generated by appending a
5176 switch specific suffix to the source file name, and the file is
5177 created in the same directory as the output file. If the
5178 @samp{-@var{options}} form is used, @var{options} is a list of
5179 @samp{-} separated options that control the details of the dump. Not
5180 all options are applicable to all dumps, those which are not
5181 meaningful will be ignored. The following options are available
5185 Print the address of each node. Usually this is not meaningful as it
5186 changes according to the environment and source file. Its primary use
5187 is for tying up a dump file with a debug environment.
5189 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
5190 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
5191 use working backward from mangled names in the assembly file.
5193 Inhibit dumping of members of a scope or body of a function merely
5194 because that scope has been reached. Only dump such items when they
5195 are directly reachable by some other path. When dumping pretty-printed
5196 trees, this option inhibits dumping the bodies of control structures.
5198 Print a raw representation of the tree. By default, trees are
5199 pretty-printed into a C-like representation.
5201 Enable more detailed dumps (not honored by every dump option).
5203 Enable dumping various statistics about the pass (not honored by every dump
5206 Enable showing basic block boundaries (disabled in raw dumps).
5208 Enable showing virtual operands for every statement.
5210 Enable showing line numbers for statements.
5212 Enable showing the unique ID (@code{DECL_UID}) for each variable.
5214 Enable showing the tree dump for each statement.
5216 Enable showing the EH region number holding each statement.
5218 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
5219 and @option{lineno}.
5222 The following tree dumps are possible:
5226 @opindex fdump-tree-original
5227 Dump before any tree based optimization, to @file{@var{file}.original}.
5230 @opindex fdump-tree-optimized
5231 Dump after all tree based optimization, to @file{@var{file}.optimized}.
5234 @opindex fdump-tree-gimple
5235 Dump each function before and after the gimplification pass to a file. The
5236 file name is made by appending @file{.gimple} to the source file name.
5239 @opindex fdump-tree-cfg
5240 Dump the control flow graph of each function to a file. The file name is
5241 made by appending @file{.cfg} to the source file name.
5244 @opindex fdump-tree-vcg
5245 Dump the control flow graph of each function to a file in VCG format. The
5246 file name is made by appending @file{.vcg} to the source file name. Note
5247 that if the file contains more than one function, the generated file cannot
5248 be used directly by VCG@. You will need to cut and paste each function's
5249 graph into its own separate file first.
5252 @opindex fdump-tree-ch
5253 Dump each function after copying loop headers. The file name is made by
5254 appending @file{.ch} to the source file name.
5257 @opindex fdump-tree-ssa
5258 Dump SSA related information to a file. The file name is made by appending
5259 @file{.ssa} to the source file name.
5262 @opindex fdump-tree-alias
5263 Dump aliasing information for each function. The file name is made by
5264 appending @file{.alias} to the source file name.
5267 @opindex fdump-tree-ccp
5268 Dump each function after CCP@. The file name is made by appending
5269 @file{.ccp} to the source file name.
5272 @opindex fdump-tree-storeccp
5273 Dump each function after STORE-CCP@. The file name is made by appending
5274 @file{.storeccp} to the source file name.
5277 @opindex fdump-tree-pre
5278 Dump trees after partial redundancy elimination. The file name is made
5279 by appending @file{.pre} to the source file name.
5282 @opindex fdump-tree-fre
5283 Dump trees after full redundancy elimination. The file name is made
5284 by appending @file{.fre} to the source file name.
5287 @opindex fdump-tree-copyprop
5288 Dump trees after copy propagation. The file name is made
5289 by appending @file{.copyprop} to the source file name.
5291 @item store_copyprop
5292 @opindex fdump-tree-store_copyprop
5293 Dump trees after store copy-propagation. The file name is made
5294 by appending @file{.store_copyprop} to the source file name.
5297 @opindex fdump-tree-dce
5298 Dump each function after dead code elimination. The file name is made by
5299 appending @file{.dce} to the source file name.
5302 @opindex fdump-tree-mudflap
5303 Dump each function after adding mudflap instrumentation. The file name is
5304 made by appending @file{.mudflap} to the source file name.
5307 @opindex fdump-tree-sra
5308 Dump each function after performing scalar replacement of aggregates. The
5309 file name is made by appending @file{.sra} to the source file name.
5312 @opindex fdump-tree-sink
5313 Dump each function after performing code sinking. The file name is made
5314 by appending @file{.sink} to the source file name.
5317 @opindex fdump-tree-dom
5318 Dump each function after applying dominator tree optimizations. The file
5319 name is made by appending @file{.dom} to the source file name.
5322 @opindex fdump-tree-dse
5323 Dump each function after applying dead store elimination. The file
5324 name is made by appending @file{.dse} to the source file name.
5327 @opindex fdump-tree-phiopt
5328 Dump each function after optimizing PHI nodes into straightline code. The file
5329 name is made by appending @file{.phiopt} to the source file name.
5332 @opindex fdump-tree-forwprop
5333 Dump each function after forward propagating single use variables. The file
5334 name is made by appending @file{.forwprop} to the source file name.
5337 @opindex fdump-tree-copyrename
5338 Dump each function after applying the copy rename optimization. The file
5339 name is made by appending @file{.copyrename} to the source file name.
5342 @opindex fdump-tree-nrv
5343 Dump each function after applying the named return value optimization on
5344 generic trees. The file name is made by appending @file{.nrv} to the source
5348 @opindex fdump-tree-vect
5349 Dump each function after applying vectorization of loops. The file name is
5350 made by appending @file{.vect} to the source file name.
5353 @opindex fdump-tree-vrp
5354 Dump each function after Value Range Propagation (VRP). The file name
5355 is made by appending @file{.vrp} to the source file name.
5358 @opindex fdump-tree-all
5359 Enable all the available tree dumps with the flags provided in this option.
5362 @item -ftree-vectorizer-verbose=@var{n}
5363 @opindex ftree-vectorizer-verbose
5364 This option controls the amount of debugging output the vectorizer prints.
5365 This information is written to standard error, unless
5366 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
5367 in which case it is output to the usual dump listing file, @file{.vect}.
5368 For @var{n}=0 no diagnostic information is reported.
5369 If @var{n}=1 the vectorizer reports each loop that got vectorized,
5370 and the total number of loops that got vectorized.
5371 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
5372 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
5373 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
5374 level that @option{-fdump-tree-vect-stats} uses.
5375 Higher verbosity levels mean either more information dumped for each
5376 reported loop, or same amount of information reported for more loops:
5377 If @var{n}=3, alignment related information is added to the reports.
5378 If @var{n}=4, data-references related information (e.g.@: memory dependences,
5379 memory access-patterns) is added to the reports.
5380 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
5381 that did not pass the first analysis phase (i.e., may not be countable, or
5382 may have complicated control-flow).
5383 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
5384 For @var{n}=7, all the information the vectorizer generates during its
5385 analysis and transformation is reported. This is the same verbosity level
5386 that @option{-fdump-tree-vect-details} uses.
5388 @item -frandom-seed=@var{string}
5389 @opindex frandom-seed
5390 This option provides a seed that GCC uses when it would otherwise use
5391 random numbers. It is used to generate certain symbol names
5392 that have to be different in every compiled file. It is also used to
5393 place unique stamps in coverage data files and the object files that
5394 produce them. You can use the @option{-frandom-seed} option to produce
5395 reproducibly identical object files.
5397 The @var{string} should be different for every file you compile.
5399 @item -fsched-verbose=@var{n}
5400 @opindex fsched-verbose
5401 On targets that use instruction scheduling, this option controls the
5402 amount of debugging output the scheduler prints. This information is
5403 written to standard error, unless @option{-fdump-rtl-sched1} or
5404 @option{-fdump-rtl-sched2} is specified, in which case it is output
5405 to the usual dump listing file, @file{.sched} or @file{.sched2}
5406 respectively. However for @var{n} greater than nine, the output is
5407 always printed to standard error.
5409 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
5410 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
5411 For @var{n} greater than one, it also output basic block probabilities,
5412 detailed ready list information and unit/insn info. For @var{n} greater
5413 than two, it includes RTL at abort point, control-flow and regions info.
5414 And for @var{n} over four, @option{-fsched-verbose} also includes
5418 @itemx -save-temps=cwd
5420 Store the usual ``temporary'' intermediate files permanently; place them
5421 in the current directory and name them based on the source file. Thus,
5422 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
5423 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
5424 preprocessed @file{foo.i} output file even though the compiler now
5425 normally uses an integrated preprocessor.
5427 When used in combination with the @option{-x} command line option,
5428 @option{-save-temps} is sensible enough to avoid over writing an
5429 input source file with the same extension as an intermediate file.
5430 The corresponding intermediate file may be obtained by renaming the
5431 source file before using @option{-save-temps}.
5433 If you invoke GCC in parallel, compiling several different source
5434 files that share a common base name in different subdirectories or the
5435 same source file compiled for multiple output destinations, it is
5436 likely that the different parallel compilers will interfere with each
5437 other, and overwrite the temporary files. For instance:
5440 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
5441 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
5444 may result in @file{foo.i} and @file{foo.o} being written to
5445 simultaneously by both compilers.
5447 @item -save-temps=obj
5448 @opindex save-temps=obj
5449 Store the usual ``temporary'' intermediate files permanently. If the
5450 @option{-o} option is used, the temporary files are based on the
5451 object file. If the @option{-o} option is not used, the
5452 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
5457 gcc -save-temps=obj -c foo.c
5458 gcc -save-temps=obj -c bar.c -o dir/xbar.o
5459 gcc -save-temps=obj foobar.c -o dir2/yfoobar
5462 would create @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
5463 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
5464 @file{dir2/yfoobar.o}.
5466 @item -time@r{[}=@var{file}@r{]}
5468 Report the CPU time taken by each subprocess in the compilation
5469 sequence. For C source files, this is the compiler proper and assembler
5470 (plus the linker if linking is done).
5472 Without the specification of an output file, the output looks like this:
5479 The first number on each line is the ``user time'', that is time spent
5480 executing the program itself. The second number is ``system time'',
5481 time spent executing operating system routines on behalf of the program.
5482 Both numbers are in seconds.
5484 With the specification of an output file, the output is appended to the
5485 named file, and it looks like this:
5488 0.12 0.01 cc1 @var{options}
5489 0.00 0.01 as @var{options}
5492 The ``user time'' and the ``system time'' are moved before the program
5493 name, and the options passed to the program are displayed, so that one
5494 can later tell what file was being compiled, and with which options.
5496 @item -fvar-tracking
5497 @opindex fvar-tracking
5498 Run variable tracking pass. It computes where variables are stored at each
5499 position in code. Better debugging information is then generated
5500 (if the debugging information format supports this information).
5502 It is enabled by default when compiling with optimization (@option{-Os},
5503 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5504 the debug info format supports it.
5506 @item -fvar-tracking-assignments
5507 @opindex fvar-tracking-assignments
5508 @opindex fno-var-tracking-assignments
5509 Annotate assignments to user variables early in the compilation and
5510 attempt to carry the annotations over throughout the compilation all the
5511 way to the end, in an attempt to improve debug information while
5512 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
5514 It can be enabled even if var-tracking is disabled, in which case
5515 annotations will be created and maintained, but discarded at the end.
5517 @item -fvar-tracking-assignments-toggle
5518 @opindex fvar-tracking-assignments-toggle
5519 @opindex fno-var-tracking-assignments-toggle
5520 Toggle @option{-fvar-tracking-assignments}, in the same way that
5521 @option{-gtoggle} toggles @option{-g}.
5523 @item -print-file-name=@var{library}
5524 @opindex print-file-name
5525 Print the full absolute name of the library file @var{library} that
5526 would be used when linking---and don't do anything else. With this
5527 option, GCC does not compile or link anything; it just prints the
5530 @item -print-multi-directory
5531 @opindex print-multi-directory
5532 Print the directory name corresponding to the multilib selected by any
5533 other switches present in the command line. This directory is supposed
5534 to exist in @env{GCC_EXEC_PREFIX}.
5536 @item -print-multi-lib
5537 @opindex print-multi-lib
5538 Print the mapping from multilib directory names to compiler switches
5539 that enable them. The directory name is separated from the switches by
5540 @samp{;}, and each switch starts with an @samp{@@} instead of the
5541 @samp{-}, without spaces between multiple switches. This is supposed to
5542 ease shell-processing.
5544 @item -print-multi-os-directory
5545 @opindex print-multi-os-directory
5546 Print the path to OS libraries for the selected
5547 multilib, relative to some @file{lib} subdirectory. If OS libraries are
5548 present in the @file{lib} subdirectory and no multilibs are used, this is
5549 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
5550 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
5551 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
5552 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
5554 @item -print-prog-name=@var{program}
5555 @opindex print-prog-name
5556 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5558 @item -print-libgcc-file-name
5559 @opindex print-libgcc-file-name
5560 Same as @option{-print-file-name=libgcc.a}.
5562 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5563 but you do want to link with @file{libgcc.a}. You can do
5566 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5569 @item -print-search-dirs
5570 @opindex print-search-dirs
5571 Print the name of the configured installation directory and a list of
5572 program and library directories @command{gcc} will search---and don't do anything else.
5574 This is useful when @command{gcc} prints the error message
5575 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5576 To resolve this you either need to put @file{cpp0} and the other compiler
5577 components where @command{gcc} expects to find them, or you can set the environment
5578 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5579 Don't forget the trailing @samp{/}.
5580 @xref{Environment Variables}.
5582 @item -print-sysroot
5583 @opindex print-sysroot
5584 Print the target sysroot directory that will be used during
5585 compilation. This is the target sysroot specified either at configure
5586 time or using the @option{--sysroot} option, possibly with an extra
5587 suffix that depends on compilation options. If no target sysroot is
5588 specified, the option prints nothing.
5590 @item -print-sysroot-headers-suffix
5591 @opindex print-sysroot-headers-suffix
5592 Print the suffix added to the target sysroot when searching for
5593 headers, or give an error if the compiler is not configured with such
5594 a suffix---and don't do anything else.
5597 @opindex dumpmachine
5598 Print the compiler's target machine (for example,
5599 @samp{i686-pc-linux-gnu})---and don't do anything else.
5602 @opindex dumpversion
5603 Print the compiler version (for example, @samp{3.0})---and don't do
5608 Print the compiler's built-in specs---and don't do anything else. (This
5609 is used when GCC itself is being built.) @xref{Spec Files}.
5611 @item -feliminate-unused-debug-types
5612 @opindex feliminate-unused-debug-types
5613 Normally, when producing DWARF2 output, GCC will emit debugging
5614 information for all types declared in a compilation
5615 unit, regardless of whether or not they are actually used
5616 in that compilation unit. Sometimes this is useful, such as
5617 if, in the debugger, you want to cast a value to a type that is
5618 not actually used in your program (but is declared). More often,
5619 however, this results in a significant amount of wasted space.
5620 With this option, GCC will avoid producing debug symbol output
5621 for types that are nowhere used in the source file being compiled.
5624 @node Optimize Options
5625 @section Options That Control Optimization
5626 @cindex optimize options
5627 @cindex options, optimization
5629 These options control various sorts of optimizations.
5631 Without any optimization option, the compiler's goal is to reduce the
5632 cost of compilation and to make debugging produce the expected
5633 results. Statements are independent: if you stop the program with a
5634 breakpoint between statements, you can then assign a new value to any
5635 variable or change the program counter to any other statement in the
5636 function and get exactly the results you would expect from the source
5639 Turning on optimization flags makes the compiler attempt to improve
5640 the performance and/or code size at the expense of compilation time
5641 and possibly the ability to debug the program.
5643 The compiler performs optimization based on the knowledge it has of the
5644 program. Compiling multiple files at once to a single output file mode allows
5645 the compiler to use information gained from all of the files when compiling
5648 Not all optimizations are controlled directly by a flag. Only
5649 optimizations that have a flag are listed in this section.
5651 Most of the optimizations are not enabled if a @option{-O} level is not set on
5652 the command line, even if individual optimization flags are specified.
5654 Depending on the target and how GCC was configured, a slightly different
5655 set of optimizations may be enabled at each @option{-O} level than
5656 those listed here. You can invoke GCC with @samp{-Q --help=optimizers}
5657 to find out the exact set of optimizations that are enabled at each level.
5658 @xref{Overall Options}, for examples.
5665 Optimize. Optimizing compilation takes somewhat more time, and a lot
5666 more memory for a large function.
5668 With @option{-O}, the compiler tries to reduce code size and execution
5669 time, without performing any optimizations that take a great deal of
5672 @option{-O} turns on the following optimization flags:
5675 -fcprop-registers @gol
5678 -fdelayed-branch @gol
5680 -fguess-branch-probability @gol
5681 -fif-conversion2 @gol
5682 -fif-conversion @gol
5683 -fipa-pure-const @gol
5684 -fipa-reference @gol
5686 -fsplit-wide-types @gol
5687 -ftree-builtin-call-dce @gol
5690 -ftree-copyrename @gol
5692 -ftree-dominator-opts @gol
5694 -ftree-forwprop @gol
5702 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5703 where doing so does not interfere with debugging.
5707 Optimize even more. GCC performs nearly all supported optimizations
5708 that do not involve a space-speed tradeoff.
5709 As compared to @option{-O}, this option increases both compilation time
5710 and the performance of the generated code.
5712 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5713 also turns on the following optimization flags:
5714 @gccoptlist{-fthread-jumps @gol
5715 -falign-functions -falign-jumps @gol
5716 -falign-loops -falign-labels @gol
5719 -fcse-follow-jumps -fcse-skip-blocks @gol
5720 -fdelete-null-pointer-checks @gol
5721 -fexpensive-optimizations @gol
5722 -fgcse -fgcse-lm @gol
5723 -finline-small-functions @gol
5724 -findirect-inlining @gol
5726 -foptimize-sibling-calls @gol
5729 -freorder-blocks -freorder-functions @gol
5730 -frerun-cse-after-loop @gol
5731 -fsched-interblock -fsched-spec @gol
5732 -fschedule-insns -fschedule-insns2 @gol
5733 -fstrict-aliasing -fstrict-overflow @gol
5734 -ftree-switch-conversion @gol
5738 Please note the warning under @option{-fgcse} about
5739 invoking @option{-O2} on programs that use computed gotos.
5743 Optimize yet more. @option{-O3} turns on all optimizations specified
5744 by @option{-O2} and also turns on the @option{-finline-functions},
5745 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5746 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5750 Reduce compilation time and make debugging produce the expected
5751 results. This is the default.
5755 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5756 do not typically increase code size. It also performs further
5757 optimizations designed to reduce code size.
5759 @option{-Os} disables the following optimization flags:
5760 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5761 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5762 -fprefetch-loop-arrays -ftree-vect-loop-version}
5764 If you use multiple @option{-O} options, with or without level numbers,
5765 the last such option is the one that is effective.
5768 Options of the form @option{-f@var{flag}} specify machine-independent
5769 flags. Most flags have both positive and negative forms; the negative
5770 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5771 below, only one of the forms is listed---the one you typically will
5772 use. You can figure out the other form by either removing @samp{no-}
5775 The following options control specific optimizations. They are either
5776 activated by @option{-O} options or are related to ones that are. You
5777 can use the following flags in the rare cases when ``fine-tuning'' of
5778 optimizations to be performed is desired.
5781 @item -fno-default-inline
5782 @opindex fno-default-inline
5783 Do not make member functions inline by default merely because they are
5784 defined inside the class scope (C++ only). Otherwise, when you specify
5785 @w{@option{-O}}, member functions defined inside class scope are compiled
5786 inline by default; i.e., you don't need to add @samp{inline} in front of
5787 the member function name.
5789 @item -fno-defer-pop
5790 @opindex fno-defer-pop
5791 Always pop the arguments to each function call as soon as that function
5792 returns. For machines which must pop arguments after a function call,
5793 the compiler normally lets arguments accumulate on the stack for several
5794 function calls and pops them all at once.
5796 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5798 @item -fforward-propagate
5799 @opindex fforward-propagate
5800 Perform a forward propagation pass on RTL@. The pass tries to combine two
5801 instructions and checks if the result can be simplified. If loop unrolling
5802 is active, two passes are performed and the second is scheduled after
5805 This option is enabled by default at optimization levels @option{-O},
5806 @option{-O2}, @option{-O3}, @option{-Os}.
5808 @item -fomit-frame-pointer
5809 @opindex fomit-frame-pointer
5810 Don't keep the frame pointer in a register for functions that
5811 don't need one. This avoids the instructions to save, set up and
5812 restore frame pointers; it also makes an extra register available
5813 in many functions. @strong{It also makes debugging impossible on
5816 On some machines, such as the VAX, this flag has no effect, because
5817 the standard calling sequence automatically handles the frame pointer
5818 and nothing is saved by pretending it doesn't exist. The
5819 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5820 whether a target machine supports this flag. @xref{Registers,,Register
5821 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5823 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5825 @item -foptimize-sibling-calls
5826 @opindex foptimize-sibling-calls
5827 Optimize sibling and tail recursive calls.
5829 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5833 Don't pay attention to the @code{inline} keyword. Normally this option
5834 is used to keep the compiler from expanding any functions inline.
5835 Note that if you are not optimizing, no functions can be expanded inline.
5837 @item -finline-small-functions
5838 @opindex finline-small-functions
5839 Integrate functions into their callers when their body is smaller than expected
5840 function call code (so overall size of program gets smaller). The compiler
5841 heuristically decides which functions are simple enough to be worth integrating
5844 Enabled at level @option{-O2}.
5846 @item -findirect-inlining
5847 @opindex findirect-inlining
5848 Inline also indirect calls that are discovered to be known at compile
5849 time thanks to previous inlining. This option has any effect only
5850 when inlining itself is turned on by the @option{-finline-functions}
5851 or @option{-finline-small-functions} options.
5853 Enabled at level @option{-O2}.
5855 @item -finline-functions
5856 @opindex finline-functions
5857 Integrate all simple functions into their callers. The compiler
5858 heuristically decides which functions are simple enough to be worth
5859 integrating in this way.
5861 If all calls to a given function are integrated, and the function is
5862 declared @code{static}, then the function is normally not output as
5863 assembler code in its own right.
5865 Enabled at level @option{-O3}.
5867 @item -finline-functions-called-once
5868 @opindex finline-functions-called-once
5869 Consider all @code{static} functions called once for inlining into their
5870 caller even if they are not marked @code{inline}. If a call to a given
5871 function is integrated, then the function is not output as assembler code
5874 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
5876 @item -fearly-inlining
5877 @opindex fearly-inlining
5878 Inline functions marked by @code{always_inline} and functions whose body seems
5879 smaller than the function call overhead early before doing
5880 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5881 makes profiling significantly cheaper and usually inlining faster on programs
5882 having large chains of nested wrapper functions.
5888 Perform interprocedural scalar replacement of aggregates, removal of
5889 unused parameters and replacement of parameters passed by reference
5890 by parameters passed by value.
5892 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
5894 @item -finline-limit=@var{n}
5895 @opindex finline-limit
5896 By default, GCC limits the size of functions that can be inlined. This flag
5897 allows coarse control of this limit. @var{n} is the size of functions that
5898 can be inlined in number of pseudo instructions.
5900 Inlining is actually controlled by a number of parameters, which may be
5901 specified individually by using @option{--param @var{name}=@var{value}}.
5902 The @option{-finline-limit=@var{n}} option sets some of these parameters
5906 @item max-inline-insns-single
5907 is set to @var{n}/2.
5908 @item max-inline-insns-auto
5909 is set to @var{n}/2.
5912 See below for a documentation of the individual
5913 parameters controlling inlining and for the defaults of these parameters.
5915 @emph{Note:} there may be no value to @option{-finline-limit} that results
5916 in default behavior.
5918 @emph{Note:} pseudo instruction represents, in this particular context, an
5919 abstract measurement of function's size. In no way does it represent a count
5920 of assembly instructions and as such its exact meaning might change from one
5921 release to an another.
5923 @item -fkeep-inline-functions
5924 @opindex fkeep-inline-functions
5925 In C, emit @code{static} functions that are declared @code{inline}
5926 into the object file, even if the function has been inlined into all
5927 of its callers. This switch does not affect functions using the
5928 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5929 inline functions into the object file.
5931 @item -fkeep-static-consts
5932 @opindex fkeep-static-consts
5933 Emit variables declared @code{static const} when optimization isn't turned
5934 on, even if the variables aren't referenced.
5936 GCC enables this option by default. If you want to force the compiler to
5937 check if the variable was referenced, regardless of whether or not
5938 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5940 @item -fmerge-constants
5941 @opindex fmerge-constants
5942 Attempt to merge identical constants (string constants and floating point
5943 constants) across compilation units.
5945 This option is the default for optimized compilation if the assembler and
5946 linker support it. Use @option{-fno-merge-constants} to inhibit this
5949 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5951 @item -fmerge-all-constants
5952 @opindex fmerge-all-constants
5953 Attempt to merge identical constants and identical variables.
5955 This option implies @option{-fmerge-constants}. In addition to
5956 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5957 arrays or initialized constant variables with integral or floating point
5958 types. Languages like C or C++ require each variable, including multiple
5959 instances of the same variable in recursive calls, to have distinct locations,
5960 so using this option will result in non-conforming
5963 @item -fmodulo-sched
5964 @opindex fmodulo-sched
5965 Perform swing modulo scheduling immediately before the first scheduling
5966 pass. This pass looks at innermost loops and reorders their
5967 instructions by overlapping different iterations.
5969 @item -fmodulo-sched-allow-regmoves
5970 @opindex fmodulo-sched-allow-regmoves
5971 Perform more aggressive SMS based modulo scheduling with register moves
5972 allowed. By setting this flag certain anti-dependences edges will be
5973 deleted which will trigger the generation of reg-moves based on the
5974 life-range analysis. This option is effective only with
5975 @option{-fmodulo-sched} enabled.
5977 @item -fno-branch-count-reg
5978 @opindex fno-branch-count-reg
5979 Do not use ``decrement and branch'' instructions on a count register,
5980 but instead generate a sequence of instructions that decrement a
5981 register, compare it against zero, then branch based upon the result.
5982 This option is only meaningful on architectures that support such
5983 instructions, which include x86, PowerPC, IA-64 and S/390.
5985 The default is @option{-fbranch-count-reg}.
5987 @item -fno-function-cse
5988 @opindex fno-function-cse
5989 Do not put function addresses in registers; make each instruction that
5990 calls a constant function contain the function's address explicitly.
5992 This option results in less efficient code, but some strange hacks
5993 that alter the assembler output may be confused by the optimizations
5994 performed when this option is not used.
5996 The default is @option{-ffunction-cse}
5998 @item -fno-zero-initialized-in-bss
5999 @opindex fno-zero-initialized-in-bss
6000 If the target supports a BSS section, GCC by default puts variables that
6001 are initialized to zero into BSS@. This can save space in the resulting
6004 This option turns off this behavior because some programs explicitly
6005 rely on variables going to the data section. E.g., so that the
6006 resulting executable can find the beginning of that section and/or make
6007 assumptions based on that.
6009 The default is @option{-fzero-initialized-in-bss}.
6011 @item -fmudflap -fmudflapth -fmudflapir
6015 @cindex bounds checking
6017 For front-ends that support it (C and C++), instrument all risky
6018 pointer/array dereferencing operations, some standard library
6019 string/heap functions, and some other associated constructs with
6020 range/validity tests. Modules so instrumented should be immune to
6021 buffer overflows, invalid heap use, and some other classes of C/C++
6022 programming errors. The instrumentation relies on a separate runtime
6023 library (@file{libmudflap}), which will be linked into a program if
6024 @option{-fmudflap} is given at link time. Run-time behavior of the
6025 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
6026 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
6029 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
6030 link if your program is multi-threaded. Use @option{-fmudflapir}, in
6031 addition to @option{-fmudflap} or @option{-fmudflapth}, if
6032 instrumentation should ignore pointer reads. This produces less
6033 instrumentation (and therefore faster execution) and still provides
6034 some protection against outright memory corrupting writes, but allows
6035 erroneously read data to propagate within a program.
6037 @item -fthread-jumps
6038 @opindex fthread-jumps
6039 Perform optimizations where we check to see if a jump branches to a
6040 location where another comparison subsumed by the first is found. If
6041 so, the first branch is redirected to either the destination of the
6042 second branch or a point immediately following it, depending on whether
6043 the condition is known to be true or false.
6045 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6047 @item -fsplit-wide-types
6048 @opindex fsplit-wide-types
6049 When using a type that occupies multiple registers, such as @code{long
6050 long} on a 32-bit system, split the registers apart and allocate them
6051 independently. This normally generates better code for those types,
6052 but may make debugging more difficult.
6054 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
6057 @item -fcse-follow-jumps
6058 @opindex fcse-follow-jumps
6059 In common subexpression elimination (CSE), scan through jump instructions
6060 when the target of the jump is not reached by any other path. For
6061 example, when CSE encounters an @code{if} statement with an
6062 @code{else} clause, CSE will follow the jump when the condition
6065 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6067 @item -fcse-skip-blocks
6068 @opindex fcse-skip-blocks
6069 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
6070 follow jumps which conditionally skip over blocks. When CSE
6071 encounters a simple @code{if} statement with no else clause,
6072 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
6073 body of the @code{if}.
6075 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6077 @item -frerun-cse-after-loop
6078 @opindex frerun-cse-after-loop
6079 Re-run common subexpression elimination after loop optimizations has been
6082 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6086 Perform a global common subexpression elimination pass.
6087 This pass also performs global constant and copy propagation.
6089 @emph{Note:} When compiling a program using computed gotos, a GCC
6090 extension, you may get better runtime performance if you disable
6091 the global common subexpression elimination pass by adding
6092 @option{-fno-gcse} to the command line.
6094 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6098 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
6099 attempt to move loads which are only killed by stores into themselves. This
6100 allows a loop containing a load/store sequence to be changed to a load outside
6101 the loop, and a copy/store within the loop.
6103 Enabled by default when gcse is enabled.
6107 When @option{-fgcse-sm} is enabled, a store motion pass is run after
6108 global common subexpression elimination. This pass will attempt to move
6109 stores out of loops. When used in conjunction with @option{-fgcse-lm},
6110 loops containing a load/store sequence can be changed to a load before
6111 the loop and a store after the loop.
6113 Not enabled at any optimization level.
6117 When @option{-fgcse-las} is enabled, the global common subexpression
6118 elimination pass eliminates redundant loads that come after stores to the
6119 same memory location (both partial and full redundancies).
6121 Not enabled at any optimization level.
6123 @item -fgcse-after-reload
6124 @opindex fgcse-after-reload
6125 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
6126 pass is performed after reload. The purpose of this pass is to cleanup
6129 @item -funsafe-loop-optimizations
6130 @opindex funsafe-loop-optimizations
6131 If given, the loop optimizer will assume that loop indices do not
6132 overflow, and that the loops with nontrivial exit condition are not
6133 infinite. This enables a wider range of loop optimizations even if
6134 the loop optimizer itself cannot prove that these assumptions are valid.
6135 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
6136 if it finds this kind of loop.
6138 @item -fcrossjumping
6139 @opindex fcrossjumping
6140 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
6141 resulting code may or may not perform better than without cross-jumping.
6143 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6145 @item -fauto-inc-dec
6146 @opindex fauto-inc-dec
6147 Combine increments or decrements of addresses with memory accesses.
6148 This pass is always skipped on architectures that do not have
6149 instructions to support this. Enabled by default at @option{-O} and
6150 higher on architectures that support this.
6154 Perform dead code elimination (DCE) on RTL@.
6155 Enabled by default at @option{-O} and higher.
6159 Perform dead store elimination (DSE) on RTL@.
6160 Enabled by default at @option{-O} and higher.
6162 @item -fif-conversion
6163 @opindex fif-conversion
6164 Attempt to transform conditional jumps into branch-less equivalents. This
6165 include use of conditional moves, min, max, set flags and abs instructions, and
6166 some tricks doable by standard arithmetics. The use of conditional execution
6167 on chips where it is available is controlled by @code{if-conversion2}.
6169 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6171 @item -fif-conversion2
6172 @opindex fif-conversion2
6173 Use conditional execution (where available) to transform conditional jumps into
6174 branch-less equivalents.
6176 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6178 @item -fdelete-null-pointer-checks
6179 @opindex fdelete-null-pointer-checks
6180 Assume that programs cannot safely dereference null pointers, and that
6181 no code or data element resides there. This enables simple constant
6182 folding optimizations at all optimization levels. In addition, other
6183 optimization passes in GCC use this flag to control global dataflow
6184 analyses that eliminate useless checks for null pointers; these assume
6185 that if a pointer is checked after it has already been dereferenced,
6188 Note however that in some environments this assumption is not true.
6189 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
6190 for programs which depend on that behavior.
6192 Some targets, especially embedded ones, disable this option at all levels.
6193 Otherwise it is enabled at all levels: @option{-O0}, @option{-O1},
6194 @option{-O2}, @option{-O3}, @option{-Os}. Passes that use the information
6195 are enabled independently at different optimization levels.
6197 @item -fexpensive-optimizations
6198 @opindex fexpensive-optimizations
6199 Perform a number of minor optimizations that are relatively expensive.
6201 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6203 @item -foptimize-register-move
6205 @opindex foptimize-register-move
6207 Attempt to reassign register numbers in move instructions and as
6208 operands of other simple instructions in order to maximize the amount of
6209 register tying. This is especially helpful on machines with two-operand
6212 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
6215 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6217 @item -fira-algorithm=@var{algorithm}
6218 Use specified coloring algorithm for the integrated register
6219 allocator. The @var{algorithm} argument should be @code{priority} or
6220 @code{CB}. The first algorithm specifies Chow's priority coloring,
6221 the second one specifies Chaitin-Briggs coloring. The second
6222 algorithm can be unimplemented for some architectures. If it is
6223 implemented, it is the default because Chaitin-Briggs coloring as a
6224 rule generates a better code.
6226 @item -fira-region=@var{region}
6227 Use specified regions for the integrated register allocator. The
6228 @var{region} argument should be one of @code{all}, @code{mixed}, or
6229 @code{one}. The first value means using all loops as register
6230 allocation regions, the second value which is the default means using
6231 all loops except for loops with small register pressure as the
6232 regions, and third one means using all function as a single region.
6233 The first value can give best result for machines with small size and
6234 irregular register set, the third one results in faster and generates
6235 decent code and the smallest size code, and the default value usually
6236 give the best results in most cases and for most architectures.
6238 @item -fira-coalesce
6239 @opindex fira-coalesce
6240 Do optimistic register coalescing. This option might be profitable for
6241 architectures with big regular register files.
6243 @item -fira-loop-pressure
6244 @opindex fira-loop-pressure
6245 Use IRA to evaluate register pressure in loops for decision to move
6246 loop invariants. Usage of this option usually results in generation
6247 of faster and smaller code on machines with big register files (>= 32
6248 registers) but it can slow compiler down.
6250 This option is enabled at level @option{-O3} for some targets.
6252 @item -fno-ira-share-save-slots
6253 @opindex fno-ira-share-save-slots
6254 Switch off sharing stack slots used for saving call used hard
6255 registers living through a call. Each hard register will get a
6256 separate stack slot and as a result function stack frame will be
6259 @item -fno-ira-share-spill-slots
6260 @opindex fno-ira-share-spill-slots
6261 Switch off sharing stack slots allocated for pseudo-registers. Each
6262 pseudo-register which did not get a hard register will get a separate
6263 stack slot and as a result function stack frame will be bigger.
6265 @item -fira-verbose=@var{n}
6266 @opindex fira-verbose
6267 Set up how verbose dump file for the integrated register allocator
6268 will be. Default value is 5. If the value is greater or equal to 10,
6269 the dump file will be stderr as if the value were @var{n} minus 10.
6271 @item -fdelayed-branch
6272 @opindex fdelayed-branch
6273 If supported for the target machine, attempt to reorder instructions
6274 to exploit instruction slots available after delayed branch
6277 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6279 @item -fschedule-insns
6280 @opindex fschedule-insns
6281 If supported for the target machine, attempt to reorder instructions to
6282 eliminate execution stalls due to required data being unavailable. This
6283 helps machines that have slow floating point or memory load instructions
6284 by allowing other instructions to be issued until the result of the load
6285 or floating point instruction is required.
6287 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6289 @item -fschedule-insns2
6290 @opindex fschedule-insns2
6291 Similar to @option{-fschedule-insns}, but requests an additional pass of
6292 instruction scheduling after register allocation has been done. This is
6293 especially useful on machines with a relatively small number of
6294 registers and where memory load instructions take more than one cycle.
6296 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6298 @item -fno-sched-interblock
6299 @opindex fno-sched-interblock
6300 Don't schedule instructions across basic blocks. This is normally
6301 enabled by default when scheduling before register allocation, i.e.@:
6302 with @option{-fschedule-insns} or at @option{-O2} or higher.
6304 @item -fno-sched-spec
6305 @opindex fno-sched-spec
6306 Don't allow speculative motion of non-load instructions. This is normally
6307 enabled by default when scheduling before register allocation, i.e.@:
6308 with @option{-fschedule-insns} or at @option{-O2} or higher.
6310 @item -fsched-pressure
6311 @opindex fsched-pressure
6312 Enable register pressure sensitive insn scheduling before the register
6313 allocation. This only makes sense when scheduling before register
6314 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
6315 @option{-O2} or higher. Usage of this option can improve the
6316 generated code and decrease its size by preventing register pressure
6317 increase above the number of available hard registers and as a
6318 consequence register spills in the register allocation.
6320 @item -fsched-spec-load
6321 @opindex fsched-spec-load
6322 Allow speculative motion of some load instructions. This only makes
6323 sense when scheduling before register allocation, i.e.@: with
6324 @option{-fschedule-insns} or at @option{-O2} or higher.
6326 @item -fsched-spec-load-dangerous
6327 @opindex fsched-spec-load-dangerous
6328 Allow speculative motion of more load instructions. This only makes
6329 sense when scheduling before register allocation, i.e.@: with
6330 @option{-fschedule-insns} or at @option{-O2} or higher.
6332 @item -fsched-stalled-insns
6333 @itemx -fsched-stalled-insns=@var{n}
6334 @opindex fsched-stalled-insns
6335 Define how many insns (if any) can be moved prematurely from the queue
6336 of stalled insns into the ready list, during the second scheduling pass.
6337 @option{-fno-sched-stalled-insns} means that no insns will be moved
6338 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
6339 on how many queued insns can be moved prematurely.
6340 @option{-fsched-stalled-insns} without a value is equivalent to
6341 @option{-fsched-stalled-insns=1}.
6343 @item -fsched-stalled-insns-dep
6344 @itemx -fsched-stalled-insns-dep=@var{n}
6345 @opindex fsched-stalled-insns-dep
6346 Define how many insn groups (cycles) will be examined for a dependency
6347 on a stalled insn that is candidate for premature removal from the queue
6348 of stalled insns. This has an effect only during the second scheduling pass,
6349 and only if @option{-fsched-stalled-insns} is used.
6350 @option{-fno-sched-stalled-insns-dep} is equivalent to
6351 @option{-fsched-stalled-insns-dep=0}.
6352 @option{-fsched-stalled-insns-dep} without a value is equivalent to
6353 @option{-fsched-stalled-insns-dep=1}.
6355 @item -fsched2-use-superblocks
6356 @opindex fsched2-use-superblocks
6357 When scheduling after register allocation, do use superblock scheduling
6358 algorithm. Superblock scheduling allows motion across basic block boundaries
6359 resulting on faster schedules. This option is experimental, as not all machine
6360 descriptions used by GCC model the CPU closely enough to avoid unreliable
6361 results from the algorithm.
6363 This only makes sense when scheduling after register allocation, i.e.@: with
6364 @option{-fschedule-insns2} or at @option{-O2} or higher.
6366 @item -fsched-group-heuristic
6367 @opindex fsched-group-heuristic
6368 Enable the group heuristic in the scheduler. This heuristic favors
6369 the instruction that belongs to a schedule group. This is enabled
6370 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6371 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6373 @item -fsched-critical-path-heuristic
6374 @opindex fsched-critical-path-heuristic
6375 Enable the critical-path heuristic in the scheduler. This heuristic favors
6376 instructions on the critical path. This is enabled by default when
6377 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6378 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6380 @item -fsched-spec-insn-heuristic
6381 @opindex fsched-spec-insn-heuristic
6382 Enable the speculative instruction heuristic in the scheduler. This
6383 heuristic favors speculative instructions with greater dependency weakness.
6384 This is enabled by default when scheduling is enabled, i.e.@:
6385 with @option{-fschedule-insns} or @option{-fschedule-insns2}
6386 or at @option{-O2} or higher.
6388 @item -fsched-rank-heuristic
6389 @opindex fsched-rank-heuristic
6390 Enable the rank heuristic in the scheduler. This heuristic favors
6391 the instruction belonging to a basic block with greater size or frequency.
6392 This is enabled by default when scheduling is enabled, i.e.@:
6393 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6394 at @option{-O2} or higher.
6396 @item -fsched-last-insn-heuristic
6397 @opindex fsched-last-insn-heuristic
6398 Enable the last-instruction heuristic in the scheduler. This heuristic
6399 favors the instruction that is less dependent on the last instruction
6400 scheduled. This is enabled by default when scheduling is enabled,
6401 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6402 at @option{-O2} or higher.
6404 @item -fsched-dep-count-heuristic
6405 @opindex fsched-dep-count-heuristic
6406 Enable the dependent-count heuristic in the scheduler. This heuristic
6407 favors the instruction that has more instructions depending on it.
6408 This is enabled by default when scheduling is enabled, i.e.@:
6409 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6410 at @option{-O2} or higher.
6412 @item -fsched2-use-traces
6413 @opindex fsched2-use-traces
6414 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
6415 allocation and additionally perform code duplication in order to increase the
6416 size of superblocks using tracer pass. See @option{-ftracer} for details on
6419 This mode should produce faster but significantly longer programs. Also
6420 without @option{-fbranch-probabilities} the traces constructed may not
6421 match the reality and hurt the performance. This only makes
6422 sense when scheduling after register allocation, i.e.@: with
6423 @option{-fschedule-insns2} or at @option{-O2} or higher.
6425 @item -freschedule-modulo-scheduled-loops
6426 @opindex freschedule-modulo-scheduled-loops
6427 The modulo scheduling comes before the traditional scheduling, if a loop
6428 was modulo scheduled we may want to prevent the later scheduling passes
6429 from changing its schedule, we use this option to control that.
6431 @item -fselective-scheduling
6432 @opindex fselective-scheduling
6433 Schedule instructions using selective scheduling algorithm. Selective
6434 scheduling runs instead of the first scheduler pass.
6436 @item -fselective-scheduling2
6437 @opindex fselective-scheduling2
6438 Schedule instructions using selective scheduling algorithm. Selective
6439 scheduling runs instead of the second scheduler pass.
6441 @item -fsel-sched-pipelining
6442 @opindex fsel-sched-pipelining
6443 Enable software pipelining of innermost loops during selective scheduling.
6444 This option has no effect until one of @option{-fselective-scheduling} or
6445 @option{-fselective-scheduling2} is turned on.
6447 @item -fsel-sched-pipelining-outer-loops
6448 @opindex fsel-sched-pipelining-outer-loops
6449 When pipelining loops during selective scheduling, also pipeline outer loops.
6450 This option has no effect until @option{-fsel-sched-pipelining} is turned on.
6452 @item -fcaller-saves
6453 @opindex fcaller-saves
6454 Enable values to be allocated in registers that will be clobbered by
6455 function calls, by emitting extra instructions to save and restore the
6456 registers around such calls. Such allocation is done only when it
6457 seems to result in better code than would otherwise be produced.
6459 This option is always enabled by default on certain machines, usually
6460 those which have no call-preserved registers to use instead.
6462 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6464 @item -fconserve-stack
6465 @opindex fconserve-stack
6466 Attempt to minimize stack usage. The compiler will attempt to use less
6467 stack space, even if that makes the program slower. This option
6468 implies setting the @option{large-stack-frame} parameter to 100
6469 and the @option{large-stack-frame-growth} parameter to 400.
6471 @item -ftree-reassoc
6472 @opindex ftree-reassoc
6473 Perform reassociation on trees. This flag is enabled by default
6474 at @option{-O} and higher.
6478 Perform partial redundancy elimination (PRE) on trees. This flag is
6479 enabled by default at @option{-O2} and @option{-O3}.
6481 @item -ftree-forwprop
6482 @opindex ftree-forwprop
6483 Perform forward propagation on trees. This flag is enabled by default
6484 at @option{-O} and higher.
6488 Perform full redundancy elimination (FRE) on trees. The difference
6489 between FRE and PRE is that FRE only considers expressions
6490 that are computed on all paths leading to the redundant computation.
6491 This analysis is faster than PRE, though it exposes fewer redundancies.
6492 This flag is enabled by default at @option{-O} and higher.
6494 @item -ftree-phiprop
6495 @opindex ftree-phiprop
6496 Perform hoisting of loads from conditional pointers on trees. This
6497 pass is enabled by default at @option{-O} and higher.
6499 @item -ftree-copy-prop
6500 @opindex ftree-copy-prop
6501 Perform copy propagation on trees. This pass eliminates unnecessary
6502 copy operations. This flag is enabled by default at @option{-O} and
6505 @item -fipa-pure-const
6506 @opindex fipa-pure-const
6507 Discover which functions are pure or constant.
6508 Enabled by default at @option{-O} and higher.
6510 @item -fipa-reference
6511 @opindex fipa-reference
6512 Discover which static variables do not escape cannot escape the
6514 Enabled by default at @option{-O} and higher.
6516 @item -fipa-struct-reorg
6517 @opindex fipa-struct-reorg
6518 Perform structure reorganization optimization, that change C-like structures
6519 layout in order to better utilize spatial locality. This transformation is
6520 affective for programs containing arrays of structures. Available in two
6521 compilation modes: profile-based (enabled with @option{-fprofile-generate})
6522 or static (which uses built-in heuristics). Require @option{-fipa-type-escape}
6523 to provide the safety of this transformation. It works only in whole program
6524 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
6525 enabled. Structures considered @samp{cold} by this transformation are not
6526 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
6528 With this flag, the program debug info reflects a new structure layout.
6532 Perform interprocedural pointer analysis. This option is experimental
6533 and does not affect generated code.
6537 Perform interprocedural constant propagation.
6538 This optimization analyzes the program to determine when values passed
6539 to functions are constants and then optimizes accordingly.
6540 This optimization can substantially increase performance
6541 if the application has constants passed to functions.
6542 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
6544 @item -fipa-cp-clone
6545 @opindex fipa-cp-clone
6546 Perform function cloning to make interprocedural constant propagation stronger.
6547 When enabled, interprocedural constant propagation will perform function cloning
6548 when externally visible function can be called with constant arguments.
6549 Because this optimization can create multiple copies of functions,
6550 it may significantly increase code size
6551 (see @option{--param ipcp-unit-growth=@var{value}}).
6552 This flag is enabled by default at @option{-O3}.
6554 @item -fipa-matrix-reorg
6555 @opindex fipa-matrix-reorg
6556 Perform matrix flattening and transposing.
6557 Matrix flattening tries to replace an @math{m}-dimensional matrix
6558 with its equivalent @math{n}-dimensional matrix, where @math{n < m}.
6559 This reduces the level of indirection needed for accessing the elements
6560 of the matrix. The second optimization is matrix transposing that
6561 attempts to change the order of the matrix's dimensions in order to
6562 improve cache locality.
6563 Both optimizations need the @option{-fwhole-program} flag.
6564 Transposing is enabled only if profiling information is available.
6568 Perform forward store motion on trees. This flag is
6569 enabled by default at @option{-O} and higher.
6573 Perform sparse conditional constant propagation (CCP) on trees. This
6574 pass only operates on local scalar variables and is enabled by default
6575 at @option{-O} and higher.
6577 @item -ftree-switch-conversion
6578 Perform conversion of simple initializations in a switch to
6579 initializations from a scalar array. This flag is enabled by default
6580 at @option{-O2} and higher.
6584 Perform dead code elimination (DCE) on trees. This flag is enabled by
6585 default at @option{-O} and higher.
6587 @item -ftree-builtin-call-dce
6588 @opindex ftree-builtin-call-dce
6589 Perform conditional dead code elimination (DCE) for calls to builtin functions
6590 that may set @code{errno} but are otherwise side-effect free. This flag is
6591 enabled by default at @option{-O2} and higher if @option{-Os} is not also
6594 @item -ftree-dominator-opts
6595 @opindex ftree-dominator-opts
6596 Perform a variety of simple scalar cleanups (constant/copy
6597 propagation, redundancy elimination, range propagation and expression
6598 simplification) based on a dominator tree traversal. This also
6599 performs jump threading (to reduce jumps to jumps). This flag is
6600 enabled by default at @option{-O} and higher.
6604 Perform dead store elimination (DSE) on trees. A dead store is a store into
6605 a memory location which will later be overwritten by another store without
6606 any intervening loads. In this case the earlier store can be deleted. This
6607 flag is enabled by default at @option{-O} and higher.
6611 Perform loop header copying on trees. This is beneficial since it increases
6612 effectiveness of code motion optimizations. It also saves one jump. This flag
6613 is enabled by default at @option{-O} and higher. It is not enabled
6614 for @option{-Os}, since it usually increases code size.
6616 @item -ftree-loop-optimize
6617 @opindex ftree-loop-optimize
6618 Perform loop optimizations on trees. This flag is enabled by default
6619 at @option{-O} and higher.
6621 @item -ftree-loop-linear
6622 @opindex ftree-loop-linear
6623 Perform linear loop transformations on tree. This flag can improve cache
6624 performance and allow further loop optimizations to take place.
6626 @item -floop-interchange
6627 Perform loop interchange transformations on loops. Interchanging two
6628 nested loops switches the inner and outer loops. For example, given a
6633 A(J, I) = A(J, I) * C
6637 loop interchange will transform the loop as if the user had written:
6641 A(J, I) = A(J, I) * C
6645 which can be beneficial when @code{N} is larger than the caches,
6646 because in Fortran, the elements of an array are stored in memory
6647 contiguously by column, and the original loop iterates over rows,
6648 potentially creating at each access a cache miss. This optimization
6649 applies to all the languages supported by GCC and is not limited to
6650 Fortran. To use this code transformation, GCC has to be configured
6651 with @option{--with-ppl} and @option{--with-cloog} to enable the
6652 Graphite loop transformation infrastructure.
6654 @item -floop-strip-mine
6655 Perform loop strip mining transformations on loops. Strip mining
6656 splits a loop into two nested loops. The outer loop has strides
6657 equal to the strip size and the inner loop has strides of the
6658 original loop within a strip. For example, given a loop like:
6664 loop strip mining will transform the loop as if the user had written:
6667 DO I = II, min (II + 3, N)
6672 This optimization applies to all the languages supported by GCC and is
6673 not limited to Fortran. To use this code transformation, GCC has to
6674 be configured with @option{--with-ppl} and @option{--with-cloog} to
6675 enable the Graphite loop transformation infrastructure.
6678 Perform loop blocking transformations on loops. Blocking strip mines
6679 each loop in the loop nest such that the memory accesses of the
6680 element loops fit inside caches. For example, given a loop like:
6684 A(J, I) = B(I) + C(J)
6688 loop blocking will transform the loop as if the user had written:
6692 DO I = II, min (II + 63, N)
6693 DO J = JJ, min (JJ + 63, M)
6694 A(J, I) = B(I) + C(J)
6700 which can be beneficial when @code{M} is larger than the caches,
6701 because the innermost loop will iterate over a smaller amount of data
6702 that can be kept in the caches. This optimization applies to all the
6703 languages supported by GCC and is not limited to Fortran. To use this
6704 code transformation, GCC has to be configured with @option{--with-ppl}
6705 and @option{--with-cloog} to enable the Graphite loop transformation
6708 @item -fgraphite-identity
6709 @opindex fgraphite-identity
6710 Enable the identity transformation for graphite. For every SCoP we generate
6711 the polyhedral representation and transform it back to gimple. Using
6712 @option{-fgraphite-identity} we can check the costs or benefits of the
6713 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
6714 are also performed by the code generator CLooG, like index splitting and
6715 dead code elimination in loops.
6717 @item -floop-parallelize-all
6718 Use the Graphite data dependence analysis to identify loops that can
6719 be parallelized. Parallelize all the loops that can be analyzed to
6720 not contain loop carried dependences without checking that it is
6721 profitable to parallelize the loops.
6723 @item -fcheck-data-deps
6724 @opindex fcheck-data-deps
6725 Compare the results of several data dependence analyzers. This option
6726 is used for debugging the data dependence analyzers.
6728 @item -ftree-loop-distribution
6729 Perform loop distribution. This flag can improve cache performance on
6730 big loop bodies and allow further loop optimizations, like
6731 parallelization or vectorization, to take place. For example, the loop
6748 @item -ftree-loop-im
6749 @opindex ftree-loop-im
6750 Perform loop invariant motion on trees. This pass moves only invariants that
6751 would be hard to handle at RTL level (function calls, operations that expand to
6752 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
6753 operands of conditions that are invariant out of the loop, so that we can use
6754 just trivial invariantness analysis in loop unswitching. The pass also includes
6757 @item -ftree-loop-ivcanon
6758 @opindex ftree-loop-ivcanon
6759 Create a canonical counter for number of iterations in the loop for that
6760 determining number of iterations requires complicated analysis. Later
6761 optimizations then may determine the number easily. Useful especially
6762 in connection with unrolling.
6766 Perform induction variable optimizations (strength reduction, induction
6767 variable merging and induction variable elimination) on trees.
6769 @item -ftree-parallelize-loops=n
6770 @opindex ftree-parallelize-loops
6771 Parallelize loops, i.e., split their iteration space to run in n threads.
6772 This is only possible for loops whose iterations are independent
6773 and can be arbitrarily reordered. The optimization is only
6774 profitable on multiprocessor machines, for loops that are CPU-intensive,
6775 rather than constrained e.g.@: by memory bandwidth. This option
6776 implies @option{-pthread}, and thus is only supported on targets
6777 that have support for @option{-pthread}.
6781 Perform function-local points-to analysis on trees. This flag is
6782 enabled by default at @option{-O} and higher.
6786 Perform scalar replacement of aggregates. This pass replaces structure
6787 references with scalars to prevent committing structures to memory too
6788 early. This flag is enabled by default at @option{-O} and higher.
6790 @item -ftree-copyrename
6791 @opindex ftree-copyrename
6792 Perform copy renaming on trees. This pass attempts to rename compiler
6793 temporaries to other variables at copy locations, usually resulting in
6794 variable names which more closely resemble the original variables. This flag
6795 is enabled by default at @option{-O} and higher.
6799 Perform temporary expression replacement during the SSA->normal phase. Single
6800 use/single def temporaries are replaced at their use location with their
6801 defining expression. This results in non-GIMPLE code, but gives the expanders
6802 much more complex trees to work on resulting in better RTL generation. This is
6803 enabled by default at @option{-O} and higher.
6805 @item -ftree-vectorize
6806 @opindex ftree-vectorize
6807 Perform loop vectorization on trees. This flag is enabled by default at
6810 @item -ftree-vect-loop-version
6811 @opindex ftree-vect-loop-version
6812 Perform loop versioning when doing loop vectorization on trees. When a loop
6813 appears to be vectorizable except that data alignment or data dependence cannot
6814 be determined at compile time then vectorized and non-vectorized versions of
6815 the loop are generated along with runtime checks for alignment or dependence
6816 to control which version is executed. This option is enabled by default
6817 except at level @option{-Os} where it is disabled.
6819 @item -fvect-cost-model
6820 @opindex fvect-cost-model
6821 Enable cost model for vectorization.
6825 Perform Value Range Propagation on trees. This is similar to the
6826 constant propagation pass, but instead of values, ranges of values are
6827 propagated. This allows the optimizers to remove unnecessary range
6828 checks like array bound checks and null pointer checks. This is
6829 enabled by default at @option{-O2} and higher. Null pointer check
6830 elimination is only done if @option{-fdelete-null-pointer-checks} is
6835 Perform tail duplication to enlarge superblock size. This transformation
6836 simplifies the control flow of the function allowing other optimizations to do
6839 @item -funroll-loops
6840 @opindex funroll-loops
6841 Unroll loops whose number of iterations can be determined at compile
6842 time or upon entry to the loop. @option{-funroll-loops} implies
6843 @option{-frerun-cse-after-loop}. This option makes code larger,
6844 and may or may not make it run faster.
6846 @item -funroll-all-loops
6847 @opindex funroll-all-loops
6848 Unroll all loops, even if their number of iterations is uncertain when
6849 the loop is entered. This usually makes programs run more slowly.
6850 @option{-funroll-all-loops} implies the same options as
6851 @option{-funroll-loops},
6853 @item -fsplit-ivs-in-unroller
6854 @opindex fsplit-ivs-in-unroller
6855 Enables expressing of values of induction variables in later iterations
6856 of the unrolled loop using the value in the first iteration. This breaks
6857 long dependency chains, thus improving efficiency of the scheduling passes.
6859 Combination of @option{-fweb} and CSE is often sufficient to obtain the
6860 same effect. However in cases the loop body is more complicated than
6861 a single basic block, this is not reliable. It also does not work at all
6862 on some of the architectures due to restrictions in the CSE pass.
6864 This optimization is enabled by default.
6866 @item -fvariable-expansion-in-unroller
6867 @opindex fvariable-expansion-in-unroller
6868 With this option, the compiler will create multiple copies of some
6869 local variables when unrolling a loop which can result in superior code.
6871 @item -fpredictive-commoning
6872 @opindex fpredictive-commoning
6873 Perform predictive commoning optimization, i.e., reusing computations
6874 (especially memory loads and stores) performed in previous
6875 iterations of loops.
6877 This option is enabled at level @option{-O3}.
6879 @item -fprefetch-loop-arrays
6880 @opindex fprefetch-loop-arrays
6881 If supported by the target machine, generate instructions to prefetch
6882 memory to improve the performance of loops that access large arrays.
6884 This option may generate better or worse code; results are highly
6885 dependent on the structure of loops within the source code.
6887 Disabled at level @option{-Os}.
6890 @itemx -fno-peephole2
6891 @opindex fno-peephole
6892 @opindex fno-peephole2
6893 Disable any machine-specific peephole optimizations. The difference
6894 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
6895 are implemented in the compiler; some targets use one, some use the
6896 other, a few use both.
6898 @option{-fpeephole} is enabled by default.
6899 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6901 @item -fno-guess-branch-probability
6902 @opindex fno-guess-branch-probability
6903 Do not guess branch probabilities using heuristics.
6905 GCC will use heuristics to guess branch probabilities if they are
6906 not provided by profiling feedback (@option{-fprofile-arcs}). These
6907 heuristics are based on the control flow graph. If some branch probabilities
6908 are specified by @samp{__builtin_expect}, then the heuristics will be
6909 used to guess branch probabilities for the rest of the control flow graph,
6910 taking the @samp{__builtin_expect} info into account. The interactions
6911 between the heuristics and @samp{__builtin_expect} can be complex, and in
6912 some cases, it may be useful to disable the heuristics so that the effects
6913 of @samp{__builtin_expect} are easier to understand.
6915 The default is @option{-fguess-branch-probability} at levels
6916 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6918 @item -freorder-blocks
6919 @opindex freorder-blocks
6920 Reorder basic blocks in the compiled function in order to reduce number of
6921 taken branches and improve code locality.
6923 Enabled at levels @option{-O2}, @option{-O3}.
6925 @item -freorder-blocks-and-partition
6926 @opindex freorder-blocks-and-partition
6927 In addition to reordering basic blocks in the compiled function, in order
6928 to reduce number of taken branches, partitions hot and cold basic blocks
6929 into separate sections of the assembly and .o files, to improve
6930 paging and cache locality performance.
6932 This optimization is automatically turned off in the presence of
6933 exception handling, for linkonce sections, for functions with a user-defined
6934 section attribute and on any architecture that does not support named
6937 @item -freorder-functions
6938 @opindex freorder-functions
6939 Reorder functions in the object file in order to
6940 improve code locality. This is implemented by using special
6941 subsections @code{.text.hot} for most frequently executed functions and
6942 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
6943 the linker so object file format must support named sections and linker must
6944 place them in a reasonable way.
6946 Also profile feedback must be available in to make this option effective. See
6947 @option{-fprofile-arcs} for details.
6949 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6951 @item -fstrict-aliasing
6952 @opindex fstrict-aliasing
6953 Allow the compiler to assume the strictest aliasing rules applicable to
6954 the language being compiled. For C (and C++), this activates
6955 optimizations based on the type of expressions. In particular, an
6956 object of one type is assumed never to reside at the same address as an
6957 object of a different type, unless the types are almost the same. For
6958 example, an @code{unsigned int} can alias an @code{int}, but not a
6959 @code{void*} or a @code{double}. A character type may alias any other
6962 @anchor{Type-punning}Pay special attention to code like this:
6975 The practice of reading from a different union member than the one most
6976 recently written to (called ``type-punning'') is common. Even with
6977 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
6978 is accessed through the union type. So, the code above will work as
6979 expected. @xref{Structures unions enumerations and bit-fields
6980 implementation}. However, this code might not:
6991 Similarly, access by taking the address, casting the resulting pointer
6992 and dereferencing the result has undefined behavior, even if the cast
6993 uses a union type, e.g.:
6997 return ((union a_union *) &d)->i;
7001 The @option{-fstrict-aliasing} option is enabled at levels
7002 @option{-O2}, @option{-O3}, @option{-Os}.
7004 @item -fstrict-overflow
7005 @opindex fstrict-overflow
7006 Allow the compiler to assume strict signed overflow rules, depending
7007 on the language being compiled. For C (and C++) this means that
7008 overflow when doing arithmetic with signed numbers is undefined, which
7009 means that the compiler may assume that it will not happen. This
7010 permits various optimizations. For example, the compiler will assume
7011 that an expression like @code{i + 10 > i} will always be true for
7012 signed @code{i}. This assumption is only valid if signed overflow is
7013 undefined, as the expression is false if @code{i + 10} overflows when
7014 using twos complement arithmetic. When this option is in effect any
7015 attempt to determine whether an operation on signed numbers will
7016 overflow must be written carefully to not actually involve overflow.
7018 This option also allows the compiler to assume strict pointer
7019 semantics: given a pointer to an object, if adding an offset to that
7020 pointer does not produce a pointer to the same object, the addition is
7021 undefined. This permits the compiler to conclude that @code{p + u >
7022 p} is always true for a pointer @code{p} and unsigned integer
7023 @code{u}. This assumption is only valid because pointer wraparound is
7024 undefined, as the expression is false if @code{p + u} overflows using
7025 twos complement arithmetic.
7027 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
7028 that integer signed overflow is fully defined: it wraps. When
7029 @option{-fwrapv} is used, there is no difference between
7030 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
7031 integers. With @option{-fwrapv} certain types of overflow are
7032 permitted. For example, if the compiler gets an overflow when doing
7033 arithmetic on constants, the overflowed value can still be used with
7034 @option{-fwrapv}, but not otherwise.
7036 The @option{-fstrict-overflow} option is enabled at levels
7037 @option{-O2}, @option{-O3}, @option{-Os}.
7039 @item -falign-functions
7040 @itemx -falign-functions=@var{n}
7041 @opindex falign-functions
7042 Align the start of functions to the next power-of-two greater than
7043 @var{n}, skipping up to @var{n} bytes. For instance,
7044 @option{-falign-functions=32} aligns functions to the next 32-byte
7045 boundary, but @option{-falign-functions=24} would align to the next
7046 32-byte boundary only if this can be done by skipping 23 bytes or less.
7048 @option{-fno-align-functions} and @option{-falign-functions=1} are
7049 equivalent and mean that functions will not be aligned.
7051 Some assemblers only support this flag when @var{n} is a power of two;
7052 in that case, it is rounded up.
7054 If @var{n} is not specified or is zero, use a machine-dependent default.
7056 Enabled at levels @option{-O2}, @option{-O3}.
7058 @item -falign-labels
7059 @itemx -falign-labels=@var{n}
7060 @opindex falign-labels
7061 Align all branch targets to a power-of-two boundary, skipping up to
7062 @var{n} bytes like @option{-falign-functions}. This option can easily
7063 make code slower, because it must insert dummy operations for when the
7064 branch target is reached in the usual flow of the code.
7066 @option{-fno-align-labels} and @option{-falign-labels=1} are
7067 equivalent and mean that labels will not be aligned.
7069 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
7070 are greater than this value, then their values are used instead.
7072 If @var{n} is not specified or is zero, use a machine-dependent default
7073 which is very likely to be @samp{1}, meaning no alignment.
7075 Enabled at levels @option{-O2}, @option{-O3}.
7078 @itemx -falign-loops=@var{n}
7079 @opindex falign-loops
7080 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
7081 like @option{-falign-functions}. The hope is that the loop will be
7082 executed many times, which will make up for any execution of the dummy
7085 @option{-fno-align-loops} and @option{-falign-loops=1} are
7086 equivalent and mean that loops will not be aligned.
7088 If @var{n} is not specified or is zero, use a machine-dependent default.
7090 Enabled at levels @option{-O2}, @option{-O3}.
7093 @itemx -falign-jumps=@var{n}
7094 @opindex falign-jumps
7095 Align branch targets to a power-of-two boundary, for branch targets
7096 where the targets can only be reached by jumping, skipping up to @var{n}
7097 bytes like @option{-falign-functions}. In this case, no dummy operations
7100 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
7101 equivalent and mean that loops will not be aligned.
7103 If @var{n} is not specified or is zero, use a machine-dependent default.
7105 Enabled at levels @option{-O2}, @option{-O3}.
7107 @item -funit-at-a-time
7108 @opindex funit-at-a-time
7109 This option is left for compatibility reasons. @option{-funit-at-a-time}
7110 has no effect, while @option{-fno-unit-at-a-time} implies
7111 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
7115 @item -fno-toplevel-reorder
7116 @opindex fno-toplevel-reorder
7117 Do not reorder top-level functions, variables, and @code{asm}
7118 statements. Output them in the same order that they appear in the
7119 input file. When this option is used, unreferenced static variables
7120 will not be removed. This option is intended to support existing code
7121 which relies on a particular ordering. For new code, it is better to
7124 Enabled at level @option{-O0}. When disabled explicitly, it also imply
7125 @option{-fno-section-anchors} that is otherwise enabled at @option{-O0} on some
7130 Constructs webs as commonly used for register allocation purposes and assign
7131 each web individual pseudo register. This allows the register allocation pass
7132 to operate on pseudos directly, but also strengthens several other optimization
7133 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
7134 however, make debugging impossible, since variables will no longer stay in a
7137 Enabled by default with @option{-funroll-loops}.
7139 @item -fwhole-program
7140 @opindex fwhole-program
7141 Assume that the current compilation unit represents the whole program being
7142 compiled. All public functions and variables with the exception of @code{main}
7143 and those merged by attribute @code{externally_visible} become static functions
7144 and in effect are optimized more aggressively by interprocedural optimizers.
7145 While this option is equivalent to proper use of the @code{static} keyword for
7146 programs consisting of a single file, in combination with option
7147 @option{-combine}, @option{-flto} or @option{-fwhopr} this flag can be used to
7148 compile many smaller scale programs since the functions and variables become
7149 local for the whole combined compilation unit, not for the single source file
7152 This option implies @option{-fwhole-file} for Fortran programs.
7156 This option runs the standard link-time optimizer. When invoked
7157 with source code, it generates GIMPLE (one of GCC's internal
7158 representations) and writes it to special ELF sections in the object
7159 file. When the object files are linked together, all the function
7160 bodies are read from these ELF sections and instantiated as if they
7161 had been part of the same translation unit.
7163 To use the link-timer optimizer, @option{-flto} needs to be specified at
7164 compile time and during the final link. For example,
7167 gcc -c -O2 -flto foo.c
7168 gcc -c -O2 -flto bar.c
7169 gcc -o myprog -flto -O2 foo.o bar.o
7172 The first two invocations to GCC will save a bytecode representation
7173 of GIMPLE into special ELF sections inside @file{foo.o} and
7174 @file{bar.o}. The final invocation will read the GIMPLE bytecode from
7175 @file{foo.o} and @file{bar.o}, merge the two files into a single
7176 internal image, and compile the result as usual. Since both
7177 @file{foo.o} and @file{bar.o} are merged into a single image, this
7178 causes all the inter-procedural analyses and optimizations in GCC to
7179 work across the two files as if they were a single one. This means,
7180 for example, that the inliner will be able to inline functions in
7181 @file{bar.o} into functions in @file{foo.o} and vice-versa.
7183 Another (simpler) way to enable link-time optimization is,
7186 gcc -o myprog -flto -O2 foo.c bar.c
7189 The above will generate bytecode for @file{foo.c} and @file{bar.c},
7190 merge them together into a single GIMPLE representation and optimize
7191 them as usual to produce @file{myprog}.
7193 The only important thing to keep in mind is that to enable link-time
7194 optimizations the @option{-flto} flag needs to be passed to both the
7195 compile and the link commands.
7197 Note that when a file is compiled with @option{-flto}, the generated
7198 object file will be larger than a regular object file because it will
7199 contain GIMPLE bytecodes and the usual final code. This means that
7200 object files with LTO information can be linked as a normal object
7201 file. So, in the previous example, if the final link is done with
7204 gcc -o myprog foo.o bar.o
7207 The only difference will be that no inter-procedural optimizations
7208 will be applied to produce @file{myprog}. The two object files
7209 @file{foo.o} and @file{bar.o} will be simply sent to the regular
7212 Additionally, the optimization flags used to compile individual files
7213 are not necessarily related to those used at link-time. For instance,
7216 gcc -c -O0 -flto foo.c
7217 gcc -c -O0 -flto bar.c
7218 gcc -o myprog -flto -O3 foo.o bar.o
7221 This will produce individual object files with unoptimized assembler
7222 code, but the resulting binary @file{myprog} will be optimized at
7223 @option{-O3}. Now, if the final binary is generated without
7224 @option{-flto}, then @file{myprog} will not be optimized.
7226 When producing the final binary with @option{-flto}, GCC will only
7227 apply link-time optimizations to those files that contain bytecode.
7228 Therefore, you can mix and match object files and libraries with
7229 GIMPLE bytecodes and final object code. GCC will automatically select
7230 which files to optimize in LTO mode and which files to link without
7233 There are some code generation flags that GCC will preserve when
7234 generating bytecodes, as they need to be used during the final link
7235 stage. Currently, the following options are saved into the GIMPLE
7236 bytecode files: @option{-fPIC}, @option{-fcommon} and all the
7237 @option{-m} target flags.
7239 At link time, these options are read-in and reapplied. Note that the
7240 current implementation makes no attempt at recognizing conflicting
7241 values for these options. If two or more files have a conflicting
7242 value (e.g., one file is compiled with @option{-fPIC} and another
7243 isn't), the compiler will simply use the last value read from the
7244 bytecode files. It is recommended, then, that all the files
7245 participating in the same link be compiled with the same options.
7247 Another feature of LTO is that it is possible to apply interprocedural
7248 optimizations on files written in different languages. This requires
7249 some support in the language front end. Currently, the C, C++ and
7250 Fortran front ends are capable of emitting GIMPLE bytecodes, so
7251 something like this should work
7256 gfortran -c -flto baz.f90
7257 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
7260 Notice that the final link is done with @command{g++} to get the C++
7261 runtime libraries and @option{-lgfortran} is added to get the Fortran
7262 runtime libraries. In general, when mixing languages in LTO mode, you
7263 should use the same link command used when mixing languages in a
7264 regular (non-LTO) compilation. This means that if your build process
7265 was mixing languages before, all you need to add is @option{-flto} to
7266 all the compile and link commands.
7268 If object files containing GIMPLE bytecode are stored in a library
7269 archive, say @file{libfoo.a}, it is possible to extract and use them
7270 in an LTO link if you are using @command{gold} as the linker (which,
7271 in turn requires GCC to be configured with @option{--enable-gold}).
7272 To enable this feature, use the flag @option{-use-linker-plugin} at
7276 gcc -o myprog -O2 -flto -use-linker-plugin a.o b.o -lfoo
7279 With the linker plugin enabled, @command{gold} will extract the needed
7280 GIMPLE files from @file{libfoo.a} and pass them on to the running GCC
7281 to make them part of the aggregated GIMPLE image to be optimized.
7283 If you are not using @command{gold} and/or do not specify
7284 @option{-use-linker-plugin} then the objects inside @file{libfoo.a}
7285 will be extracted and linked as usual, but they will not participate
7286 in the LTO optimization process.
7288 Link time optimizations do not require the presence of the whole
7289 program to operate. If the program does not require any symbols to
7290 be exported, it is possible to combine @option{-flto} and
7291 @option{-fwhopr} with @option{-fwhole-program} to allow the
7292 interprocedural optimizers to use more aggressive assumptions which
7293 may lead to improved optimization opportunities.
7295 Regarding portability: the current implementation of LTO makes no
7296 attempt at generating bytecode that can be ported between different
7297 types of hosts. The bytecode files are versioned and there is a
7298 strict version check, so bytecode files generated in one version of
7299 GCC will not work with an older/newer version of GCC.
7301 This option is disabled by default.
7305 This option is identical in functionality to @option{-flto} but it
7306 differs in how the final link stage is executed. Instead of loading
7307 all the function bodies in memory, the callgraph is analyzed and
7308 optimization decisions are made (whole program analysis or WPA). Once
7309 optimization decisions are made, the callgraph is partitioned and the
7310 different sections are compiled separately (local transformations or
7311 LTRANS)@. This process allows optimizations on very large programs
7312 that otherwise would not fit in memory. This option enables
7313 @option{-fwpa} and @option{-fltrans} automatically.
7315 Disabled by default.
7319 This is an internal option used by GCC when compiling with
7320 @option{-fwhopr}. You should never need to use it.
7322 This option runs the link-time optimizer in the whole-program-analysis
7323 (WPA) mode, which reads in summary information from all inputs and
7324 performs a whole-program analysis based on summary information only.
7325 It generates object files for subsequent runs of the link-time
7326 optimizer where individual object files are optimized using both
7327 summary information from the WPA mode and the actual function bodies.
7328 It then drives the LTRANS phase.
7330 Disabled by default.
7334 This is an internal option used by GCC when compiling with
7335 @option{-fwhopr}. You should never need to use it.
7337 This option runs the link-time optimizer in the local-transformation (LTRANS)
7338 mode, which reads in output from a previous run of the LTO in WPA mode.
7339 In the LTRANS mode, LTO optimizes an object and produces the final assembly.
7341 Disabled by default.
7343 @item -fltrans-output-list=@var{file}
7344 @opindex fltrans-output-list
7345 This is an internal option used by GCC when compiling with
7346 @option{-fwhopr}. You should never need to use it.
7348 This option specifies a file to which the names of LTRANS output files are
7349 written. This option is only meaningful in conjunction with @option{-fwpa}.
7351 Disabled by default.
7353 @item -flto-compression-level=@var{n}
7354 This option specifies the level of compression used for intermediate
7355 language written to LTO object files, and is only meaningful in
7356 conjunction with LTO mode (@option{-fwhopr}, @option{-flto}). Valid
7357 values are 0 (no compression) to 9 (maximum compression). Values
7358 outside this range are clamped to either 0 or 9. If the option is not
7359 given, a default balanced compression setting is used.
7362 Prints a report with internal details on the workings of the link-time
7363 optimizer. The contents of this report vary from version to version,
7364 it is meant to be useful to GCC developers when processing object
7365 files in LTO mode (via @option{-fwhopr} or @option{-flto}).
7367 Disabled by default.
7369 @item -use-linker-plugin
7370 Enables the extraction of objects with GIMPLE bytecode information
7371 from library archives. This option relies on features available only
7372 in @command{gold}, so to use this you must configure GCC with
7373 @option{--enable-gold}. See @option{-flto} for a description on the
7374 effect of this flag and how to use it.
7376 Disabled by default.
7378 @item -fcprop-registers
7379 @opindex fcprop-registers
7380 After register allocation and post-register allocation instruction splitting,
7381 we perform a copy-propagation pass to try to reduce scheduling dependencies
7382 and occasionally eliminate the copy.
7384 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7386 @item -fprofile-correction
7387 @opindex fprofile-correction
7388 Profiles collected using an instrumented binary for multi-threaded programs may
7389 be inconsistent due to missed counter updates. When this option is specified,
7390 GCC will use heuristics to correct or smooth out such inconsistencies. By
7391 default, GCC will emit an error message when an inconsistent profile is detected.
7393 @item -fprofile-dir=@var{path}
7394 @opindex fprofile-dir
7396 Set the directory to search the profile data files in to @var{path}.
7397 This option affects only the profile data generated by
7398 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
7399 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
7400 and its related options.
7401 By default, GCC will use the current directory as @var{path}
7402 thus the profile data file will appear in the same directory as the object file.
7404 @item -fprofile-generate
7405 @itemx -fprofile-generate=@var{path}
7406 @opindex fprofile-generate
7408 Enable options usually used for instrumenting application to produce
7409 profile useful for later recompilation with profile feedback based
7410 optimization. You must use @option{-fprofile-generate} both when
7411 compiling and when linking your program.
7413 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
7415 If @var{path} is specified, GCC will look at the @var{path} to find
7416 the profile feedback data files. See @option{-fprofile-dir}.
7419 @itemx -fprofile-use=@var{path}
7420 @opindex fprofile-use
7421 Enable profile feedback directed optimizations, and optimizations
7422 generally profitable only with profile feedback available.
7424 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
7425 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
7427 By default, GCC emits an error message if the feedback profiles do not
7428 match the source code. This error can be turned into a warning by using
7429 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
7432 If @var{path} is specified, GCC will look at the @var{path} to find
7433 the profile feedback data files. See @option{-fprofile-dir}.
7436 The following options control compiler behavior regarding floating
7437 point arithmetic. These options trade off between speed and
7438 correctness. All must be specifically enabled.
7442 @opindex ffloat-store
7443 Do not store floating point variables in registers, and inhibit other
7444 options that might change whether a floating point value is taken from a
7447 @cindex floating point precision
7448 This option prevents undesirable excess precision on machines such as
7449 the 68000 where the floating registers (of the 68881) keep more
7450 precision than a @code{double} is supposed to have. Similarly for the
7451 x86 architecture. For most programs, the excess precision does only
7452 good, but a few programs rely on the precise definition of IEEE floating
7453 point. Use @option{-ffloat-store} for such programs, after modifying
7454 them to store all pertinent intermediate computations into variables.
7456 @item -fexcess-precision=@var{style}
7457 @opindex fexcess-precision
7458 This option allows further control over excess precision on machines
7459 where floating-point registers have more precision than the IEEE
7460 @code{float} and @code{double} types and the processor does not
7461 support operations rounding to those types. By default,
7462 @option{-fexcess-precision=fast} is in effect; this means that
7463 operations are carried out in the precision of the registers and that
7464 it is unpredictable when rounding to the types specified in the source
7465 code takes place. When compiling C, if
7466 @option{-fexcess-precision=standard} is specified then excess
7467 precision will follow the rules specified in ISO C99; in particular,
7468 both casts and assignments cause values to be rounded to their
7469 semantic types (whereas @option{-ffloat-store} only affects
7470 assignments). This option is enabled by default for C if a strict
7471 conformance option such as @option{-std=c99} is used.
7474 @option{-fexcess-precision=standard} is not implemented for languages
7475 other than C, and has no effect if
7476 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
7477 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
7478 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
7479 semantics apply without excess precision, and in the latter, rounding
7484 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
7485 @option{-ffinite-math-only}, @option{-fno-rounding-math},
7486 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
7488 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
7490 This option is not turned on by any @option{-O} option since
7491 it can result in incorrect output for programs which depend on
7492 an exact implementation of IEEE or ISO rules/specifications for
7493 math functions. It may, however, yield faster code for programs
7494 that do not require the guarantees of these specifications.
7496 @item -fno-math-errno
7497 @opindex fno-math-errno
7498 Do not set ERRNO after calling math functions that are executed
7499 with a single instruction, e.g., sqrt. A program that relies on
7500 IEEE exceptions for math error handling may want to use this flag
7501 for speed while maintaining IEEE arithmetic compatibility.
7503 This option is not turned on by any @option{-O} option since
7504 it can result in incorrect output for programs which depend on
7505 an exact implementation of IEEE or ISO rules/specifications for
7506 math functions. It may, however, yield faster code for programs
7507 that do not require the guarantees of these specifications.
7509 The default is @option{-fmath-errno}.
7511 On Darwin systems, the math library never sets @code{errno}. There is
7512 therefore no reason for the compiler to consider the possibility that
7513 it might, and @option{-fno-math-errno} is the default.
7515 @item -funsafe-math-optimizations
7516 @opindex funsafe-math-optimizations
7518 Allow optimizations for floating-point arithmetic that (a) assume
7519 that arguments and results are valid and (b) may violate IEEE or
7520 ANSI standards. When used at link-time, it may include libraries
7521 or startup files that change the default FPU control word or other
7522 similar optimizations.
7524 This option is not turned on by any @option{-O} option since
7525 it can result in incorrect output for programs which depend on
7526 an exact implementation of IEEE or ISO rules/specifications for
7527 math functions. It may, however, yield faster code for programs
7528 that do not require the guarantees of these specifications.
7529 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
7530 @option{-fassociative-math} and @option{-freciprocal-math}.
7532 The default is @option{-fno-unsafe-math-optimizations}.
7534 @item -fassociative-math
7535 @opindex fassociative-math
7537 Allow re-association of operands in series of floating-point operations.
7538 This violates the ISO C and C++ language standard by possibly changing
7539 computation result. NOTE: re-ordering may change the sign of zero as
7540 well as ignore NaNs and inhibit or create underflow or overflow (and
7541 thus cannot be used on a code which relies on rounding behavior like
7542 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
7543 and thus may not be used when ordered comparisons are required.
7544 This option requires that both @option{-fno-signed-zeros} and
7545 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
7546 much sense with @option{-frounding-math}.
7548 The default is @option{-fno-associative-math}.
7550 @item -freciprocal-math
7551 @opindex freciprocal-math
7553 Allow the reciprocal of a value to be used instead of dividing by
7554 the value if this enables optimizations. For example @code{x / y}
7555 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
7556 is subject to common subexpression elimination. Note that this loses
7557 precision and increases the number of flops operating on the value.
7559 The default is @option{-fno-reciprocal-math}.
7561 @item -ffinite-math-only
7562 @opindex ffinite-math-only
7563 Allow optimizations for floating-point arithmetic that assume
7564 that arguments and results are not NaNs or +-Infs.
7566 This option is not turned on by any @option{-O} option since
7567 it can result in incorrect output for programs which depend on
7568 an exact implementation of IEEE or ISO rules/specifications for
7569 math functions. It may, however, yield faster code for programs
7570 that do not require the guarantees of these specifications.
7572 The default is @option{-fno-finite-math-only}.
7574 @item -fno-signed-zeros
7575 @opindex fno-signed-zeros
7576 Allow optimizations for floating point arithmetic that ignore the
7577 signedness of zero. IEEE arithmetic specifies the behavior of
7578 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
7579 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
7580 This option implies that the sign of a zero result isn't significant.
7582 The default is @option{-fsigned-zeros}.
7584 @item -fno-trapping-math
7585 @opindex fno-trapping-math
7586 Compile code assuming that floating-point operations cannot generate
7587 user-visible traps. These traps include division by zero, overflow,
7588 underflow, inexact result and invalid operation. This option requires
7589 that @option{-fno-signaling-nans} be in effect. Setting this option may
7590 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
7592 This option should never be turned on by any @option{-O} option since
7593 it can result in incorrect output for programs which depend on
7594 an exact implementation of IEEE or ISO rules/specifications for
7597 The default is @option{-ftrapping-math}.
7599 @item -frounding-math
7600 @opindex frounding-math
7601 Disable transformations and optimizations that assume default floating
7602 point rounding behavior. This is round-to-zero for all floating point
7603 to integer conversions, and round-to-nearest for all other arithmetic
7604 truncations. This option should be specified for programs that change
7605 the FP rounding mode dynamically, or that may be executed with a
7606 non-default rounding mode. This option disables constant folding of
7607 floating point expressions at compile-time (which may be affected by
7608 rounding mode) and arithmetic transformations that are unsafe in the
7609 presence of sign-dependent rounding modes.
7611 The default is @option{-fno-rounding-math}.
7613 This option is experimental and does not currently guarantee to
7614 disable all GCC optimizations that are affected by rounding mode.
7615 Future versions of GCC may provide finer control of this setting
7616 using C99's @code{FENV_ACCESS} pragma. This command line option
7617 will be used to specify the default state for @code{FENV_ACCESS}.
7619 @item -fsignaling-nans
7620 @opindex fsignaling-nans
7621 Compile code assuming that IEEE signaling NaNs may generate user-visible
7622 traps during floating-point operations. Setting this option disables
7623 optimizations that may change the number of exceptions visible with
7624 signaling NaNs. This option implies @option{-ftrapping-math}.
7626 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
7629 The default is @option{-fno-signaling-nans}.
7631 This option is experimental and does not currently guarantee to
7632 disable all GCC optimizations that affect signaling NaN behavior.
7634 @item -fsingle-precision-constant
7635 @opindex fsingle-precision-constant
7636 Treat floating point constant as single precision constant instead of
7637 implicitly converting it to double precision constant.
7639 @item -fcx-limited-range
7640 @opindex fcx-limited-range
7641 When enabled, this option states that a range reduction step is not
7642 needed when performing complex division. Also, there is no checking
7643 whether the result of a complex multiplication or division is @code{NaN
7644 + I*NaN}, with an attempt to rescue the situation in that case. The
7645 default is @option{-fno-cx-limited-range}, but is enabled by
7646 @option{-ffast-math}.
7648 This option controls the default setting of the ISO C99
7649 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
7652 @item -fcx-fortran-rules
7653 @opindex fcx-fortran-rules
7654 Complex multiplication and division follow Fortran rules. Range
7655 reduction is done as part of complex division, but there is no checking
7656 whether the result of a complex multiplication or division is @code{NaN
7657 + I*NaN}, with an attempt to rescue the situation in that case.
7659 The default is @option{-fno-cx-fortran-rules}.
7663 The following options control optimizations that may improve
7664 performance, but are not enabled by any @option{-O} options. This
7665 section includes experimental options that may produce broken code.
7668 @item -fbranch-probabilities
7669 @opindex fbranch-probabilities
7670 After running a program compiled with @option{-fprofile-arcs}
7671 (@pxref{Debugging Options,, Options for Debugging Your Program or
7672 @command{gcc}}), you can compile it a second time using
7673 @option{-fbranch-probabilities}, to improve optimizations based on
7674 the number of times each branch was taken. When the program
7675 compiled with @option{-fprofile-arcs} exits it saves arc execution
7676 counts to a file called @file{@var{sourcename}.gcda} for each source
7677 file. The information in this data file is very dependent on the
7678 structure of the generated code, so you must use the same source code
7679 and the same optimization options for both compilations.
7681 With @option{-fbranch-probabilities}, GCC puts a
7682 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
7683 These can be used to improve optimization. Currently, they are only
7684 used in one place: in @file{reorg.c}, instead of guessing which path a
7685 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
7686 exactly determine which path is taken more often.
7688 @item -fprofile-values
7689 @opindex fprofile-values
7690 If combined with @option{-fprofile-arcs}, it adds code so that some
7691 data about values of expressions in the program is gathered.
7693 With @option{-fbranch-probabilities}, it reads back the data gathered
7694 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
7695 notes to instructions for their later usage in optimizations.
7697 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
7701 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
7702 a code to gather information about values of expressions.
7704 With @option{-fbranch-probabilities}, it reads back the data gathered
7705 and actually performs the optimizations based on them.
7706 Currently the optimizations include specialization of division operation
7707 using the knowledge about the value of the denominator.
7709 @item -frename-registers
7710 @opindex frename-registers
7711 Attempt to avoid false dependencies in scheduled code by making use
7712 of registers left over after register allocation. This optimization
7713 will most benefit processors with lots of registers. Depending on the
7714 debug information format adopted by the target, however, it can
7715 make debugging impossible, since variables will no longer stay in
7716 a ``home register''.
7718 Enabled by default with @option{-funroll-loops}.
7722 Perform tail duplication to enlarge superblock size. This transformation
7723 simplifies the control flow of the function allowing other optimizations to do
7726 Enabled with @option{-fprofile-use}.
7728 @item -funroll-loops
7729 @opindex funroll-loops
7730 Unroll loops whose number of iterations can be determined at compile time or
7731 upon entry to the loop. @option{-funroll-loops} implies
7732 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
7733 It also turns on complete loop peeling (i.e.@: complete removal of loops with
7734 small constant number of iterations). This option makes code larger, and may
7735 or may not make it run faster.
7737 Enabled with @option{-fprofile-use}.
7739 @item -funroll-all-loops
7740 @opindex funroll-all-loops
7741 Unroll all loops, even if their number of iterations is uncertain when
7742 the loop is entered. This usually makes programs run more slowly.
7743 @option{-funroll-all-loops} implies the same options as
7744 @option{-funroll-loops}.
7747 @opindex fpeel-loops
7748 Peels the loops for that there is enough information that they do not
7749 roll much (from profile feedback). It also turns on complete loop peeling
7750 (i.e.@: complete removal of loops with small constant number of iterations).
7752 Enabled with @option{-fprofile-use}.
7754 @item -fmove-loop-invariants
7755 @opindex fmove-loop-invariants
7756 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
7757 at level @option{-O1}
7759 @item -funswitch-loops
7760 @opindex funswitch-loops
7761 Move branches with loop invariant conditions out of the loop, with duplicates
7762 of the loop on both branches (modified according to result of the condition).
7764 @item -ffunction-sections
7765 @itemx -fdata-sections
7766 @opindex ffunction-sections
7767 @opindex fdata-sections
7768 Place each function or data item into its own section in the output
7769 file if the target supports arbitrary sections. The name of the
7770 function or the name of the data item determines the section's name
7773 Use these options on systems where the linker can perform optimizations
7774 to improve locality of reference in the instruction space. Most systems
7775 using the ELF object format and SPARC processors running Solaris 2 have
7776 linkers with such optimizations. AIX may have these optimizations in
7779 Only use these options when there are significant benefits from doing
7780 so. When you specify these options, the assembler and linker will
7781 create larger object and executable files and will also be slower.
7782 You will not be able to use @code{gprof} on all systems if you
7783 specify this option and you may have problems with debugging if
7784 you specify both this option and @option{-g}.
7786 @item -fbranch-target-load-optimize
7787 @opindex fbranch-target-load-optimize
7788 Perform branch target register load optimization before prologue / epilogue
7790 The use of target registers can typically be exposed only during reload,
7791 thus hoisting loads out of loops and doing inter-block scheduling needs
7792 a separate optimization pass.
7794 @item -fbranch-target-load-optimize2
7795 @opindex fbranch-target-load-optimize2
7796 Perform branch target register load optimization after prologue / epilogue
7799 @item -fbtr-bb-exclusive
7800 @opindex fbtr-bb-exclusive
7801 When performing branch target register load optimization, don't reuse
7802 branch target registers in within any basic block.
7804 @item -fstack-protector
7805 @opindex fstack-protector
7806 Emit extra code to check for buffer overflows, such as stack smashing
7807 attacks. This is done by adding a guard variable to functions with
7808 vulnerable objects. This includes functions that call alloca, and
7809 functions with buffers larger than 8 bytes. The guards are initialized
7810 when a function is entered and then checked when the function exits.
7811 If a guard check fails, an error message is printed and the program exits.
7813 @item -fstack-protector-all
7814 @opindex fstack-protector-all
7815 Like @option{-fstack-protector} except that all functions are protected.
7817 @item -fsection-anchors
7818 @opindex fsection-anchors
7819 Try to reduce the number of symbolic address calculations by using
7820 shared ``anchor'' symbols to address nearby objects. This transformation
7821 can help to reduce the number of GOT entries and GOT accesses on some
7824 For example, the implementation of the following function @code{foo}:
7828 int foo (void) @{ return a + b + c; @}
7831 would usually calculate the addresses of all three variables, but if you
7832 compile it with @option{-fsection-anchors}, it will access the variables
7833 from a common anchor point instead. The effect is similar to the
7834 following pseudocode (which isn't valid C):
7839 register int *xr = &x;
7840 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
7844 Not all targets support this option.
7846 @item --param @var{name}=@var{value}
7848 In some places, GCC uses various constants to control the amount of
7849 optimization that is done. For example, GCC will not inline functions
7850 that contain more that a certain number of instructions. You can
7851 control some of these constants on the command-line using the
7852 @option{--param} option.
7854 The names of specific parameters, and the meaning of the values, are
7855 tied to the internals of the compiler, and are subject to change
7856 without notice in future releases.
7858 In each case, the @var{value} is an integer. The allowable choices for
7859 @var{name} are given in the following table:
7862 @item struct-reorg-cold-struct-ratio
7863 The threshold ratio (as a percentage) between a structure frequency
7864 and the frequency of the hottest structure in the program. This parameter
7865 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
7866 We say that if the ratio of a structure frequency, calculated by profiling,
7867 to the hottest structure frequency in the program is less than this
7868 parameter, then structure reorganization is not applied to this structure.
7871 @item predictable-branch-cost-outcome
7872 When branch is predicted to be taken with probability lower than this threshold
7873 (in percent), then it is considered well predictable. The default is 10.
7875 @item max-crossjump-edges
7876 The maximum number of incoming edges to consider for crossjumping.
7877 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
7878 the number of edges incoming to each block. Increasing values mean
7879 more aggressive optimization, making the compile time increase with
7880 probably small improvement in executable size.
7882 @item min-crossjump-insns
7883 The minimum number of instructions which must be matched at the end
7884 of two blocks before crossjumping will be performed on them. This
7885 value is ignored in the case where all instructions in the block being
7886 crossjumped from are matched. The default value is 5.
7888 @item max-grow-copy-bb-insns
7889 The maximum code size expansion factor when copying basic blocks
7890 instead of jumping. The expansion is relative to a jump instruction.
7891 The default value is 8.
7893 @item max-goto-duplication-insns
7894 The maximum number of instructions to duplicate to a block that jumps
7895 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
7896 passes, GCC factors computed gotos early in the compilation process,
7897 and unfactors them as late as possible. Only computed jumps at the
7898 end of a basic blocks with no more than max-goto-duplication-insns are
7899 unfactored. The default value is 8.
7901 @item max-delay-slot-insn-search
7902 The maximum number of instructions to consider when looking for an
7903 instruction to fill a delay slot. If more than this arbitrary number of
7904 instructions is searched, the time savings from filling the delay slot
7905 will be minimal so stop searching. Increasing values mean more
7906 aggressive optimization, making the compile time increase with probably
7907 small improvement in executable run time.
7909 @item max-delay-slot-live-search
7910 When trying to fill delay slots, the maximum number of instructions to
7911 consider when searching for a block with valid live register
7912 information. Increasing this arbitrarily chosen value means more
7913 aggressive optimization, increasing the compile time. This parameter
7914 should be removed when the delay slot code is rewritten to maintain the
7917 @item max-gcse-memory
7918 The approximate maximum amount of memory that will be allocated in
7919 order to perform the global common subexpression elimination
7920 optimization. If more memory than specified is required, the
7921 optimization will not be done.
7923 @item max-pending-list-length
7924 The maximum number of pending dependencies scheduling will allow
7925 before flushing the current state and starting over. Large functions
7926 with few branches or calls can create excessively large lists which
7927 needlessly consume memory and resources.
7929 @item max-inline-insns-single
7930 Several parameters control the tree inliner used in gcc.
7931 This number sets the maximum number of instructions (counted in GCC's
7932 internal representation) in a single function that the tree inliner
7933 will consider for inlining. This only affects functions declared
7934 inline and methods implemented in a class declaration (C++).
7935 The default value is 300.
7937 @item max-inline-insns-auto
7938 When you use @option{-finline-functions} (included in @option{-O3}),
7939 a lot of functions that would otherwise not be considered for inlining
7940 by the compiler will be investigated. To those functions, a different
7941 (more restrictive) limit compared to functions declared inline can
7943 The default value is 50.
7945 @item large-function-insns
7946 The limit specifying really large functions. For functions larger than this
7947 limit after inlining, inlining is constrained by
7948 @option{--param large-function-growth}. This parameter is useful primarily
7949 to avoid extreme compilation time caused by non-linear algorithms used by the
7951 The default value is 2700.
7953 @item large-function-growth
7954 Specifies maximal growth of large function caused by inlining in percents.
7955 The default value is 100 which limits large function growth to 2.0 times
7958 @item large-unit-insns
7959 The limit specifying large translation unit. Growth caused by inlining of
7960 units larger than this limit is limited by @option{--param inline-unit-growth}.
7961 For small units this might be too tight (consider unit consisting of function A
7962 that is inline and B that just calls A three time. If B is small relative to
7963 A, the growth of unit is 300\% and yet such inlining is very sane. For very
7964 large units consisting of small inlineable functions however the overall unit
7965 growth limit is needed to avoid exponential explosion of code size. Thus for
7966 smaller units, the size is increased to @option{--param large-unit-insns}
7967 before applying @option{--param inline-unit-growth}. The default is 10000
7969 @item inline-unit-growth
7970 Specifies maximal overall growth of the compilation unit caused by inlining.
7971 The default value is 30 which limits unit growth to 1.3 times the original
7974 @item ipcp-unit-growth
7975 Specifies maximal overall growth of the compilation unit caused by
7976 interprocedural constant propagation. The default value is 10 which limits
7977 unit growth to 1.1 times the original size.
7979 @item large-stack-frame
7980 The limit specifying large stack frames. While inlining the algorithm is trying
7981 to not grow past this limit too much. Default value is 256 bytes.
7983 @item large-stack-frame-growth
7984 Specifies maximal growth of large stack frames caused by inlining in percents.
7985 The default value is 1000 which limits large stack frame growth to 11 times
7988 @item max-inline-insns-recursive
7989 @itemx max-inline-insns-recursive-auto
7990 Specifies maximum number of instructions out-of-line copy of self recursive inline
7991 function can grow into by performing recursive inlining.
7993 For functions declared inline @option{--param max-inline-insns-recursive} is
7994 taken into account. For function not declared inline, recursive inlining
7995 happens only when @option{-finline-functions} (included in @option{-O3}) is
7996 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
7997 default value is 450.
7999 @item max-inline-recursive-depth
8000 @itemx max-inline-recursive-depth-auto
8001 Specifies maximum recursion depth used by the recursive inlining.
8003 For functions declared inline @option{--param max-inline-recursive-depth} is
8004 taken into account. For function not declared inline, recursive inlining
8005 happens only when @option{-finline-functions} (included in @option{-O3}) is
8006 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
8009 @item min-inline-recursive-probability
8010 Recursive inlining is profitable only for function having deep recursion
8011 in average and can hurt for function having little recursion depth by
8012 increasing the prologue size or complexity of function body to other
8015 When profile feedback is available (see @option{-fprofile-generate}) the actual
8016 recursion depth can be guessed from probability that function will recurse via
8017 given call expression. This parameter limits inlining only to call expression
8018 whose probability exceeds given threshold (in percents). The default value is
8021 @item early-inlining-insns
8022 Specify growth that early inliner can make. In effect it increases amount of
8023 inlining for code having large abstraction penalty. The default value is 8.
8025 @item max-early-inliner-iterations
8026 @itemx max-early-inliner-iterations
8027 Limit of iterations of early inliner. This basically bounds number of nested
8028 indirect calls early inliner can resolve. Deeper chains are still handled by
8031 @item min-vect-loop-bound
8032 The minimum number of iterations under which a loop will not get vectorized
8033 when @option{-ftree-vectorize} is used. The number of iterations after
8034 vectorization needs to be greater than the value specified by this option
8035 to allow vectorization. The default value is 0.
8037 @item max-unrolled-insns
8038 The maximum number of instructions that a loop should have if that loop
8039 is unrolled, and if the loop is unrolled, it determines how many times
8040 the loop code is unrolled.
8042 @item max-average-unrolled-insns
8043 The maximum number of instructions biased by probabilities of their execution
8044 that a loop should have if that loop is unrolled, and if the loop is unrolled,
8045 it determines how many times the loop code is unrolled.
8047 @item max-unroll-times
8048 The maximum number of unrollings of a single loop.
8050 @item max-peeled-insns
8051 The maximum number of instructions that a loop should have if that loop
8052 is peeled, and if the loop is peeled, it determines how many times
8053 the loop code is peeled.
8055 @item max-peel-times
8056 The maximum number of peelings of a single loop.
8058 @item max-completely-peeled-insns
8059 The maximum number of insns of a completely peeled loop.
8061 @item max-completely-peel-times
8062 The maximum number of iterations of a loop to be suitable for complete peeling.
8064 @item max-unswitch-insns
8065 The maximum number of insns of an unswitched loop.
8067 @item max-unswitch-level
8068 The maximum number of branches unswitched in a single loop.
8071 The minimum cost of an expensive expression in the loop invariant motion.
8073 @item iv-consider-all-candidates-bound
8074 Bound on number of candidates for induction variables below that
8075 all candidates are considered for each use in induction variable
8076 optimizations. Only the most relevant candidates are considered
8077 if there are more candidates, to avoid quadratic time complexity.
8079 @item iv-max-considered-uses
8080 The induction variable optimizations give up on loops that contain more
8081 induction variable uses.
8083 @item iv-always-prune-cand-set-bound
8084 If number of candidates in the set is smaller than this value,
8085 we always try to remove unnecessary ivs from the set during its
8086 optimization when a new iv is added to the set.
8088 @item scev-max-expr-size
8089 Bound on size of expressions used in the scalar evolutions analyzer.
8090 Large expressions slow the analyzer.
8092 @item omega-max-vars
8093 The maximum number of variables in an Omega constraint system.
8094 The default value is 128.
8096 @item omega-max-geqs
8097 The maximum number of inequalities in an Omega constraint system.
8098 The default value is 256.
8101 The maximum number of equalities in an Omega constraint system.
8102 The default value is 128.
8104 @item omega-max-wild-cards
8105 The maximum number of wildcard variables that the Omega solver will
8106 be able to insert. The default value is 18.
8108 @item omega-hash-table-size
8109 The size of the hash table in the Omega solver. The default value is
8112 @item omega-max-keys
8113 The maximal number of keys used by the Omega solver. The default
8116 @item omega-eliminate-redundant-constraints
8117 When set to 1, use expensive methods to eliminate all redundant
8118 constraints. The default value is 0.
8120 @item vect-max-version-for-alignment-checks
8121 The maximum number of runtime checks that can be performed when
8122 doing loop versioning for alignment in the vectorizer. See option
8123 ftree-vect-loop-version for more information.
8125 @item vect-max-version-for-alias-checks
8126 The maximum number of runtime checks that can be performed when
8127 doing loop versioning for alias in the vectorizer. See option
8128 ftree-vect-loop-version for more information.
8130 @item max-iterations-to-track
8132 The maximum number of iterations of a loop the brute force algorithm
8133 for analysis of # of iterations of the loop tries to evaluate.
8135 @item hot-bb-count-fraction
8136 Select fraction of the maximal count of repetitions of basic block in program
8137 given basic block needs to have to be considered hot.
8139 @item hot-bb-frequency-fraction
8140 Select fraction of the maximal frequency of executions of basic block in
8141 function given basic block needs to have to be considered hot
8143 @item max-predicted-iterations
8144 The maximum number of loop iterations we predict statically. This is useful
8145 in cases where function contain single loop with known bound and other loop
8146 with unknown. We predict the known number of iterations correctly, while
8147 the unknown number of iterations average to roughly 10. This means that the
8148 loop without bounds would appear artificially cold relative to the other one.
8150 @item align-threshold
8152 Select fraction of the maximal frequency of executions of basic block in
8153 function given basic block will get aligned.
8155 @item align-loop-iterations
8157 A loop expected to iterate at lest the selected number of iterations will get
8160 @item tracer-dynamic-coverage
8161 @itemx tracer-dynamic-coverage-feedback
8163 This value is used to limit superblock formation once the given percentage of
8164 executed instructions is covered. This limits unnecessary code size
8167 The @option{tracer-dynamic-coverage-feedback} is used only when profile
8168 feedback is available. The real profiles (as opposed to statically estimated
8169 ones) are much less balanced allowing the threshold to be larger value.
8171 @item tracer-max-code-growth
8172 Stop tail duplication once code growth has reached given percentage. This is
8173 rather hokey argument, as most of the duplicates will be eliminated later in
8174 cross jumping, so it may be set to much higher values than is the desired code
8177 @item tracer-min-branch-ratio
8179 Stop reverse growth when the reverse probability of best edge is less than this
8180 threshold (in percent).
8182 @item tracer-min-branch-ratio
8183 @itemx tracer-min-branch-ratio-feedback
8185 Stop forward growth if the best edge do have probability lower than this
8188 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
8189 compilation for profile feedback and one for compilation without. The value
8190 for compilation with profile feedback needs to be more conservative (higher) in
8191 order to make tracer effective.
8193 @item max-cse-path-length
8195 Maximum number of basic blocks on path that cse considers. The default is 10.
8198 The maximum instructions CSE process before flushing. The default is 1000.
8200 @item ggc-min-expand
8202 GCC uses a garbage collector to manage its own memory allocation. This
8203 parameter specifies the minimum percentage by which the garbage
8204 collector's heap should be allowed to expand between collections.
8205 Tuning this may improve compilation speed; it has no effect on code
8208 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
8209 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
8210 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
8211 GCC is not able to calculate RAM on a particular platform, the lower
8212 bound of 30% is used. Setting this parameter and
8213 @option{ggc-min-heapsize} to zero causes a full collection to occur at
8214 every opportunity. This is extremely slow, but can be useful for
8217 @item ggc-min-heapsize
8219 Minimum size of the garbage collector's heap before it begins bothering
8220 to collect garbage. The first collection occurs after the heap expands
8221 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
8222 tuning this may improve compilation speed, and has no effect on code
8225 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
8226 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
8227 with a lower bound of 4096 (four megabytes) and an upper bound of
8228 131072 (128 megabytes). If GCC is not able to calculate RAM on a
8229 particular platform, the lower bound is used. Setting this parameter
8230 very large effectively disables garbage collection. Setting this
8231 parameter and @option{ggc-min-expand} to zero causes a full collection
8232 to occur at every opportunity.
8234 @item max-reload-search-insns
8235 The maximum number of instruction reload should look backward for equivalent
8236 register. Increasing values mean more aggressive optimization, making the
8237 compile time increase with probably slightly better performance. The default
8240 @item max-cselib-memory-locations
8241 The maximum number of memory locations cselib should take into account.
8242 Increasing values mean more aggressive optimization, making the compile time
8243 increase with probably slightly better performance. The default value is 500.
8245 @item reorder-blocks-duplicate
8246 @itemx reorder-blocks-duplicate-feedback
8248 Used by basic block reordering pass to decide whether to use unconditional
8249 branch or duplicate the code on its destination. Code is duplicated when its
8250 estimated size is smaller than this value multiplied by the estimated size of
8251 unconditional jump in the hot spots of the program.
8253 The @option{reorder-block-duplicate-feedback} is used only when profile
8254 feedback is available and may be set to higher values than
8255 @option{reorder-block-duplicate} since information about the hot spots is more
8258 @item max-sched-ready-insns
8259 The maximum number of instructions ready to be issued the scheduler should
8260 consider at any given time during the first scheduling pass. Increasing
8261 values mean more thorough searches, making the compilation time increase
8262 with probably little benefit. The default value is 100.
8264 @item max-sched-region-blocks
8265 The maximum number of blocks in a region to be considered for
8266 interblock scheduling. The default value is 10.
8268 @item max-pipeline-region-blocks
8269 The maximum number of blocks in a region to be considered for
8270 pipelining in the selective scheduler. The default value is 15.
8272 @item max-sched-region-insns
8273 The maximum number of insns in a region to be considered for
8274 interblock scheduling. The default value is 100.
8276 @item max-pipeline-region-insns
8277 The maximum number of insns in a region to be considered for
8278 pipelining in the selective scheduler. The default value is 200.
8281 The minimum probability (in percents) of reaching a source block
8282 for interblock speculative scheduling. The default value is 40.
8284 @item max-sched-extend-regions-iters
8285 The maximum number of iterations through CFG to extend regions.
8286 0 - disable region extension,
8287 N - do at most N iterations.
8288 The default value is 0.
8290 @item max-sched-insn-conflict-delay
8291 The maximum conflict delay for an insn to be considered for speculative motion.
8292 The default value is 3.
8294 @item sched-spec-prob-cutoff
8295 The minimal probability of speculation success (in percents), so that
8296 speculative insn will be scheduled.
8297 The default value is 40.
8299 @item sched-mem-true-dep-cost
8300 Minimal distance (in CPU cycles) between store and load targeting same
8301 memory locations. The default value is 1.
8303 @item selsched-max-lookahead
8304 The maximum size of the lookahead window of selective scheduling. It is a
8305 depth of search for available instructions.
8306 The default value is 50.
8308 @item selsched-max-sched-times
8309 The maximum number of times that an instruction will be scheduled during
8310 selective scheduling. This is the limit on the number of iterations
8311 through which the instruction may be pipelined. The default value is 2.
8313 @item selsched-max-insns-to-rename
8314 The maximum number of best instructions in the ready list that are considered
8315 for renaming in the selective scheduler. The default value is 2.
8317 @item max-last-value-rtl
8318 The maximum size measured as number of RTLs that can be recorded in an expression
8319 in combiner for a pseudo register as last known value of that register. The default
8322 @item integer-share-limit
8323 Small integer constants can use a shared data structure, reducing the
8324 compiler's memory usage and increasing its speed. This sets the maximum
8325 value of a shared integer constant. The default value is 256.
8327 @item min-virtual-mappings
8328 Specifies the minimum number of virtual mappings in the incremental
8329 SSA updater that should be registered to trigger the virtual mappings
8330 heuristic defined by virtual-mappings-ratio. The default value is
8333 @item virtual-mappings-ratio
8334 If the number of virtual mappings is virtual-mappings-ratio bigger
8335 than the number of virtual symbols to be updated, then the incremental
8336 SSA updater switches to a full update for those symbols. The default
8339 @item ssp-buffer-size
8340 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
8341 protection when @option{-fstack-protection} is used.
8343 @item max-jump-thread-duplication-stmts
8344 Maximum number of statements allowed in a block that needs to be
8345 duplicated when threading jumps.
8347 @item max-fields-for-field-sensitive
8348 Maximum number of fields in a structure we will treat in
8349 a field sensitive manner during pointer analysis. The default is zero
8350 for -O0, and -O1 and 100 for -Os, -O2, and -O3.
8352 @item prefetch-latency
8353 Estimate on average number of instructions that are executed before
8354 prefetch finishes. The distance we prefetch ahead is proportional
8355 to this constant. Increasing this number may also lead to less
8356 streams being prefetched (see @option{simultaneous-prefetches}).
8358 @item simultaneous-prefetches
8359 Maximum number of prefetches that can run at the same time.
8361 @item l1-cache-line-size
8362 The size of cache line in L1 cache, in bytes.
8365 The size of L1 cache, in kilobytes.
8368 The size of L2 cache, in kilobytes.
8370 @item min-insn-to-prefetch-ratio
8371 The minimum ratio between the number of instructions and the
8372 number of prefetches to enable prefetching in a loop with an
8375 @item prefetch-min-insn-to-mem-ratio
8376 The minimum ratio between the number of instructions and the
8377 number of memory references to enable prefetching in a loop.
8379 @item use-canonical-types
8380 Whether the compiler should use the ``canonical'' type system. By
8381 default, this should always be 1, which uses a more efficient internal
8382 mechanism for comparing types in C++ and Objective-C++. However, if
8383 bugs in the canonical type system are causing compilation failures,
8384 set this value to 0 to disable canonical types.
8386 @item switch-conversion-max-branch-ratio
8387 Switch initialization conversion will refuse to create arrays that are
8388 bigger than @option{switch-conversion-max-branch-ratio} times the number of
8389 branches in the switch.
8391 @item max-partial-antic-length
8392 Maximum length of the partial antic set computed during the tree
8393 partial redundancy elimination optimization (@option{-ftree-pre}) when
8394 optimizing at @option{-O3} and above. For some sorts of source code
8395 the enhanced partial redundancy elimination optimization can run away,
8396 consuming all of the memory available on the host machine. This
8397 parameter sets a limit on the length of the sets that are computed,
8398 which prevents the runaway behavior. Setting a value of 0 for
8399 this parameter will allow an unlimited set length.
8401 @item sccvn-max-scc-size
8402 Maximum size of a strongly connected component (SCC) during SCCVN
8403 processing. If this limit is hit, SCCVN processing for the whole
8404 function will not be done and optimizations depending on it will
8405 be disabled. The default maximum SCC size is 10000.
8407 @item ira-max-loops-num
8408 IRA uses a regional register allocation by default. If a function
8409 contains loops more than number given by the parameter, only at most
8410 given number of the most frequently executed loops will form regions
8411 for the regional register allocation. The default value of the
8414 @item ira-max-conflict-table-size
8415 Although IRA uses a sophisticated algorithm of compression conflict
8416 table, the table can be still big for huge functions. If the conflict
8417 table for a function could be more than size in MB given by the
8418 parameter, the conflict table is not built and faster, simpler, and
8419 lower quality register allocation algorithm will be used. The
8420 algorithm do not use pseudo-register conflicts. The default value of
8421 the parameter is 2000.
8423 @item ira-loop-reserved-regs
8424 IRA can be used to evaluate more accurate register pressure in loops
8425 for decision to move loop invariants (see @option{-O3}). The number
8426 of available registers reserved for some other purposes is described
8427 by this parameter. The default value of the parameter is 2 which is
8428 minimal number of registers needed for execution of typical
8429 instruction. This value is the best found from numerous experiments.
8431 @item loop-invariant-max-bbs-in-loop
8432 Loop invariant motion can be very expensive, both in compile time and
8433 in amount of needed compile time memory, with very large loops. Loops
8434 with more basic blocks than this parameter won't have loop invariant
8435 motion optimization performed on them. The default value of the
8436 parameter is 1000 for -O1 and 10000 for -O2 and above.
8438 @item min-nondebug-insn-uid
8439 Use uids starting at this parameter for nondebug insns. The range below
8440 the parameter is reserved exclusively for debug insns created by
8441 @option{-fvar-tracking-assignments}, but debug insns may get
8442 (non-overlapping) uids above it if the reserved range is exhausted.
8444 @item ipa-sra-ptr-growth-factor
8445 IPA-SRA will replace a pointer to an aggregate with one or more new
8446 parameters only when their cumulative size is less or equal to
8447 @option{ipa-sra-ptr-growth-factor} times the size of the original
8453 @node Preprocessor Options
8454 @section Options Controlling the Preprocessor
8455 @cindex preprocessor options
8456 @cindex options, preprocessor
8458 These options control the C preprocessor, which is run on each C source
8459 file before actual compilation.
8461 If you use the @option{-E} option, nothing is done except preprocessing.
8462 Some of these options make sense only together with @option{-E} because
8463 they cause the preprocessor output to be unsuitable for actual
8467 @item -Wp,@var{option}
8469 You can use @option{-Wp,@var{option}} to bypass the compiler driver
8470 and pass @var{option} directly through to the preprocessor. If
8471 @var{option} contains commas, it is split into multiple options at the
8472 commas. However, many options are modified, translated or interpreted
8473 by the compiler driver before being passed to the preprocessor, and
8474 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
8475 interface is undocumented and subject to change, so whenever possible
8476 you should avoid using @option{-Wp} and let the driver handle the
8479 @item -Xpreprocessor @var{option}
8480 @opindex Xpreprocessor
8481 Pass @var{option} as an option to the preprocessor. You can use this to
8482 supply system-specific preprocessor options which GCC does not know how to
8485 If you want to pass an option that takes an argument, you must use
8486 @option{-Xpreprocessor} twice, once for the option and once for the argument.
8489 @include cppopts.texi
8491 @node Assembler Options
8492 @section Passing Options to the Assembler
8494 @c prevent bad page break with this line
8495 You can pass options to the assembler.
8498 @item -Wa,@var{option}
8500 Pass @var{option} as an option to the assembler. If @var{option}
8501 contains commas, it is split into multiple options at the commas.
8503 @item -Xassembler @var{option}
8505 Pass @var{option} as an option to the assembler. You can use this to
8506 supply system-specific assembler options which GCC does not know how to
8509 If you want to pass an option that takes an argument, you must use
8510 @option{-Xassembler} twice, once for the option and once for the argument.
8515 @section Options for Linking
8516 @cindex link options
8517 @cindex options, linking
8519 These options come into play when the compiler links object files into
8520 an executable output file. They are meaningless if the compiler is
8521 not doing a link step.
8525 @item @var{object-file-name}
8526 A file name that does not end in a special recognized suffix is
8527 considered to name an object file or library. (Object files are
8528 distinguished from libraries by the linker according to the file
8529 contents.) If linking is done, these object files are used as input
8538 If any of these options is used, then the linker is not run, and
8539 object file names should not be used as arguments. @xref{Overall
8543 @item -l@var{library}
8544 @itemx -l @var{library}
8546 Search the library named @var{library} when linking. (The second
8547 alternative with the library as a separate argument is only for
8548 POSIX compliance and is not recommended.)
8550 It makes a difference where in the command you write this option; the
8551 linker searches and processes libraries and object files in the order they
8552 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
8553 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
8554 to functions in @samp{z}, those functions may not be loaded.
8556 The linker searches a standard list of directories for the library,
8557 which is actually a file named @file{lib@var{library}.a}. The linker
8558 then uses this file as if it had been specified precisely by name.
8560 The directories searched include several standard system directories
8561 plus any that you specify with @option{-L}.
8563 Normally the files found this way are library files---archive files
8564 whose members are object files. The linker handles an archive file by
8565 scanning through it for members which define symbols that have so far
8566 been referenced but not defined. But if the file that is found is an
8567 ordinary object file, it is linked in the usual fashion. The only
8568 difference between using an @option{-l} option and specifying a file name
8569 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
8570 and searches several directories.
8574 You need this special case of the @option{-l} option in order to
8575 link an Objective-C or Objective-C++ program.
8578 @opindex nostartfiles
8579 Do not use the standard system startup files when linking.
8580 The standard system libraries are used normally, unless @option{-nostdlib}
8581 or @option{-nodefaultlibs} is used.
8583 @item -nodefaultlibs
8584 @opindex nodefaultlibs
8585 Do not use the standard system libraries when linking.
8586 Only the libraries you specify will be passed to the linker, options
8587 specifying linkage of the system libraries, such as @code{-static-libgcc}
8588 or @code{-shared-libgcc}, will be ignored.
8589 The standard startup files are used normally, unless @option{-nostartfiles}
8590 is used. The compiler may generate calls to @code{memcmp},
8591 @code{memset}, @code{memcpy} and @code{memmove}.
8592 These entries are usually resolved by entries in
8593 libc. These entry points should be supplied through some other
8594 mechanism when this option is specified.
8598 Do not use the standard system startup files or libraries when linking.
8599 No startup files and only the libraries you specify will be passed to
8600 the linker, options specifying linkage of the system libraries, such as
8601 @code{-static-libgcc} or @code{-shared-libgcc}, will be ignored.
8602 The compiler may generate calls to @code{memcmp}, @code{memset},
8603 @code{memcpy} and @code{memmove}.
8604 These entries are usually resolved by entries in
8605 libc. These entry points should be supplied through some other
8606 mechanism when this option is specified.
8608 @cindex @option{-lgcc}, use with @option{-nostdlib}
8609 @cindex @option{-nostdlib} and unresolved references
8610 @cindex unresolved references and @option{-nostdlib}
8611 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
8612 @cindex @option{-nodefaultlibs} and unresolved references
8613 @cindex unresolved references and @option{-nodefaultlibs}
8614 One of the standard libraries bypassed by @option{-nostdlib} and
8615 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
8616 that GCC uses to overcome shortcomings of particular machines, or special
8617 needs for some languages.
8618 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
8619 Collection (GCC) Internals},
8620 for more discussion of @file{libgcc.a}.)
8621 In most cases, you need @file{libgcc.a} even when you want to avoid
8622 other standard libraries. In other words, when you specify @option{-nostdlib}
8623 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
8624 This ensures that you have no unresolved references to internal GCC
8625 library subroutines. (For example, @samp{__main}, used to ensure C++
8626 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
8627 GNU Compiler Collection (GCC) Internals}.)
8631 Produce a position independent executable on targets which support it.
8632 For predictable results, you must also specify the same set of options
8633 that were used to generate code (@option{-fpie}, @option{-fPIE},
8634 or model suboptions) when you specify this option.
8638 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
8639 that support it. This instructs the linker to add all symbols, not
8640 only used ones, to the dynamic symbol table. This option is needed
8641 for some uses of @code{dlopen} or to allow obtaining backtraces
8642 from within a program.
8646 Remove all symbol table and relocation information from the executable.
8650 On systems that support dynamic linking, this prevents linking with the shared
8651 libraries. On other systems, this option has no effect.
8655 Produce a shared object which can then be linked with other objects to
8656 form an executable. Not all systems support this option. For predictable
8657 results, you must also specify the same set of options that were used to
8658 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
8659 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
8660 needs to build supplementary stub code for constructors to work. On
8661 multi-libbed systems, @samp{gcc -shared} must select the correct support
8662 libraries to link against. Failing to supply the correct flags may lead
8663 to subtle defects. Supplying them in cases where they are not necessary
8666 @item -shared-libgcc
8667 @itemx -static-libgcc
8668 @opindex shared-libgcc
8669 @opindex static-libgcc
8670 On systems that provide @file{libgcc} as a shared library, these options
8671 force the use of either the shared or static version respectively.
8672 If no shared version of @file{libgcc} was built when the compiler was
8673 configured, these options have no effect.
8675 There are several situations in which an application should use the
8676 shared @file{libgcc} instead of the static version. The most common
8677 of these is when the application wishes to throw and catch exceptions
8678 across different shared libraries. In that case, each of the libraries
8679 as well as the application itself should use the shared @file{libgcc}.
8681 Therefore, the G++ and GCJ drivers automatically add
8682 @option{-shared-libgcc} whenever you build a shared library or a main
8683 executable, because C++ and Java programs typically use exceptions, so
8684 this is the right thing to do.
8686 If, instead, you use the GCC driver to create shared libraries, you may
8687 find that they will not always be linked with the shared @file{libgcc}.
8688 If GCC finds, at its configuration time, that you have a non-GNU linker
8689 or a GNU linker that does not support option @option{--eh-frame-hdr},
8690 it will link the shared version of @file{libgcc} into shared libraries
8691 by default. Otherwise, it will take advantage of the linker and optimize
8692 away the linking with the shared version of @file{libgcc}, linking with
8693 the static version of libgcc by default. This allows exceptions to
8694 propagate through such shared libraries, without incurring relocation
8695 costs at library load time.
8697 However, if a library or main executable is supposed to throw or catch
8698 exceptions, you must link it using the G++ or GCJ driver, as appropriate
8699 for the languages used in the program, or using the option
8700 @option{-shared-libgcc}, such that it is linked with the shared
8703 @item -static-libstdc++
8704 When the @command{g++} program is used to link a C++ program, it will
8705 normally automatically link against @option{libstdc++}. If
8706 @file{libstdc++} is available as a shared library, and the
8707 @option{-static} option is not used, then this will link against the
8708 shared version of @file{libstdc++}. That is normally fine. However, it
8709 is sometimes useful to freeze the version of @file{libstdc++} used by
8710 the program without going all the way to a fully static link. The
8711 @option{-static-libstdc++} option directs the @command{g++} driver to
8712 link @file{libstdc++} statically, without necessarily linking other
8713 libraries statically.
8717 Bind references to global symbols when building a shared object. Warn
8718 about any unresolved references (unless overridden by the link editor
8719 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
8722 @item -T @var{script}
8724 @cindex linker script
8725 Use @var{script} as the linker script. This option is supported by most
8726 systems using the GNU linker. On some targets, such as bare-board
8727 targets without an operating system, the @option{-T} option may be required
8728 when linking to avoid references to undefined symbols.
8730 @item -Xlinker @var{option}
8732 Pass @var{option} as an option to the linker. You can use this to
8733 supply system-specific linker options which GCC does not know how to
8736 If you want to pass an option that takes a separate argument, you must use
8737 @option{-Xlinker} twice, once for the option and once for the argument.
8738 For example, to pass @option{-assert definitions}, you must write
8739 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
8740 @option{-Xlinker "-assert definitions"}, because this passes the entire
8741 string as a single argument, which is not what the linker expects.
8743 When using the GNU linker, it is usually more convenient to pass
8744 arguments to linker options using the @option{@var{option}=@var{value}}
8745 syntax than as separate arguments. For example, you can specify
8746 @samp{-Xlinker -Map=output.map} rather than
8747 @samp{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
8748 this syntax for command-line options.
8750 @item -Wl,@var{option}
8752 Pass @var{option} as an option to the linker. If @var{option} contains
8753 commas, it is split into multiple options at the commas. You can use this
8754 syntax to pass an argument to the option.
8755 For example, @samp{-Wl,-Map,output.map} passes @samp{-Map output.map} to the
8756 linker. When using the GNU linker, you can also get the same effect with
8757 @samp{-Wl,-Map=output.map}.
8759 @item -u @var{symbol}
8761 Pretend the symbol @var{symbol} is undefined, to force linking of
8762 library modules to define it. You can use @option{-u} multiple times with
8763 different symbols to force loading of additional library modules.
8766 @node Directory Options
8767 @section Options for Directory Search
8768 @cindex directory options
8769 @cindex options, directory search
8772 These options specify directories to search for header files, for
8773 libraries and for parts of the compiler:
8778 Add the directory @var{dir} to the head of the list of directories to be
8779 searched for header files. This can be used to override a system header
8780 file, substituting your own version, since these directories are
8781 searched before the system header file directories. However, you should
8782 not use this option to add directories that contain vendor-supplied
8783 system header files (use @option{-isystem} for that). If you use more than
8784 one @option{-I} option, the directories are scanned in left-to-right
8785 order; the standard system directories come after.
8787 If a standard system include directory, or a directory specified with
8788 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
8789 option will be ignored. The directory will still be searched but as a
8790 system directory at its normal position in the system include chain.
8791 This is to ensure that GCC's procedure to fix buggy system headers and
8792 the ordering for the include_next directive are not inadvertently changed.
8793 If you really need to change the search order for system directories,
8794 use the @option{-nostdinc} and/or @option{-isystem} options.
8796 @item -iquote@var{dir}
8798 Add the directory @var{dir} to the head of the list of directories to
8799 be searched for header files only for the case of @samp{#include
8800 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
8801 otherwise just like @option{-I}.
8805 Add directory @var{dir} to the list of directories to be searched
8808 @item -B@var{prefix}
8810 This option specifies where to find the executables, libraries,
8811 include files, and data files of the compiler itself.
8813 The compiler driver program runs one or more of the subprograms
8814 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
8815 @var{prefix} as a prefix for each program it tries to run, both with and
8816 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
8818 For each subprogram to be run, the compiler driver first tries the
8819 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
8820 was not specified, the driver tries two standard prefixes, which are
8821 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
8822 those results in a file name that is found, the unmodified program
8823 name is searched for using the directories specified in your
8824 @env{PATH} environment variable.
8826 The compiler will check to see if the path provided by the @option{-B}
8827 refers to a directory, and if necessary it will add a directory
8828 separator character at the end of the path.
8830 @option{-B} prefixes that effectively specify directory names also apply
8831 to libraries in the linker, because the compiler translates these
8832 options into @option{-L} options for the linker. They also apply to
8833 includes files in the preprocessor, because the compiler translates these
8834 options into @option{-isystem} options for the preprocessor. In this case,
8835 the compiler appends @samp{include} to the prefix.
8837 The run-time support file @file{libgcc.a} can also be searched for using
8838 the @option{-B} prefix, if needed. If it is not found there, the two
8839 standard prefixes above are tried, and that is all. The file is left
8840 out of the link if it is not found by those means.
8842 Another way to specify a prefix much like the @option{-B} prefix is to use
8843 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
8846 As a special kludge, if the path provided by @option{-B} is
8847 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
8848 9, then it will be replaced by @file{[dir/]include}. This is to help
8849 with boot-strapping the compiler.
8851 @item -specs=@var{file}
8853 Process @var{file} after the compiler reads in the standard @file{specs}
8854 file, in order to override the defaults that the @file{gcc} driver
8855 program uses when determining what switches to pass to @file{cc1},
8856 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
8857 @option{-specs=@var{file}} can be specified on the command line, and they
8858 are processed in order, from left to right.
8860 @item --sysroot=@var{dir}
8862 Use @var{dir} as the logical root directory for headers and libraries.
8863 For example, if the compiler would normally search for headers in
8864 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
8865 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
8867 If you use both this option and the @option{-isysroot} option, then
8868 the @option{--sysroot} option will apply to libraries, but the
8869 @option{-isysroot} option will apply to header files.
8871 The GNU linker (beginning with version 2.16) has the necessary support
8872 for this option. If your linker does not support this option, the
8873 header file aspect of @option{--sysroot} will still work, but the
8874 library aspect will not.
8878 This option has been deprecated. Please use @option{-iquote} instead for
8879 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
8880 Any directories you specify with @option{-I} options before the @option{-I-}
8881 option are searched only for the case of @samp{#include "@var{file}"};
8882 they are not searched for @samp{#include <@var{file}>}.
8884 If additional directories are specified with @option{-I} options after
8885 the @option{-I-}, these directories are searched for all @samp{#include}
8886 directives. (Ordinarily @emph{all} @option{-I} directories are used
8889 In addition, the @option{-I-} option inhibits the use of the current
8890 directory (where the current input file came from) as the first search
8891 directory for @samp{#include "@var{file}"}. There is no way to
8892 override this effect of @option{-I-}. With @option{-I.} you can specify
8893 searching the directory which was current when the compiler was
8894 invoked. That is not exactly the same as what the preprocessor does
8895 by default, but it is often satisfactory.
8897 @option{-I-} does not inhibit the use of the standard system directories
8898 for header files. Thus, @option{-I-} and @option{-nostdinc} are
8905 @section Specifying subprocesses and the switches to pass to them
8908 @command{gcc} is a driver program. It performs its job by invoking a
8909 sequence of other programs to do the work of compiling, assembling and
8910 linking. GCC interprets its command-line parameters and uses these to
8911 deduce which programs it should invoke, and which command-line options
8912 it ought to place on their command lines. This behavior is controlled
8913 by @dfn{spec strings}. In most cases there is one spec string for each
8914 program that GCC can invoke, but a few programs have multiple spec
8915 strings to control their behavior. The spec strings built into GCC can
8916 be overridden by using the @option{-specs=} command-line switch to specify
8919 @dfn{Spec files} are plaintext files that are used to construct spec
8920 strings. They consist of a sequence of directives separated by blank
8921 lines. The type of directive is determined by the first non-whitespace
8922 character on the line and it can be one of the following:
8925 @item %@var{command}
8926 Issues a @var{command} to the spec file processor. The commands that can
8930 @item %include <@var{file}>
8932 Search for @var{file} and insert its text at the current point in the
8935 @item %include_noerr <@var{file}>
8936 @cindex %include_noerr
8937 Just like @samp{%include}, but do not generate an error message if the include
8938 file cannot be found.
8940 @item %rename @var{old_name} @var{new_name}
8942 Rename the spec string @var{old_name} to @var{new_name}.
8946 @item *[@var{spec_name}]:
8947 This tells the compiler to create, override or delete the named spec
8948 string. All lines after this directive up to the next directive or
8949 blank line are considered to be the text for the spec string. If this
8950 results in an empty string then the spec will be deleted. (Or, if the
8951 spec did not exist, then nothing will happened.) Otherwise, if the spec
8952 does not currently exist a new spec will be created. If the spec does
8953 exist then its contents will be overridden by the text of this
8954 directive, unless the first character of that text is the @samp{+}
8955 character, in which case the text will be appended to the spec.
8957 @item [@var{suffix}]:
8958 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
8959 and up to the next directive or blank line are considered to make up the
8960 spec string for the indicated suffix. When the compiler encounters an
8961 input file with the named suffix, it will processes the spec string in
8962 order to work out how to compile that file. For example:
8969 This says that any input file whose name ends in @samp{.ZZ} should be
8970 passed to the program @samp{z-compile}, which should be invoked with the
8971 command-line switch @option{-input} and with the result of performing the
8972 @samp{%i} substitution. (See below.)
8974 As an alternative to providing a spec string, the text that follows a
8975 suffix directive can be one of the following:
8978 @item @@@var{language}
8979 This says that the suffix is an alias for a known @var{language}. This is
8980 similar to using the @option{-x} command-line switch to GCC to specify a
8981 language explicitly. For example:
8988 Says that .ZZ files are, in fact, C++ source files.
8991 This causes an error messages saying:
8994 @var{name} compiler not installed on this system.
8998 GCC already has an extensive list of suffixes built into it.
8999 This directive will add an entry to the end of the list of suffixes, but
9000 since the list is searched from the end backwards, it is effectively
9001 possible to override earlier entries using this technique.
9005 GCC has the following spec strings built into it. Spec files can
9006 override these strings or create their own. Note that individual
9007 targets can also add their own spec strings to this list.
9010 asm Options to pass to the assembler
9011 asm_final Options to pass to the assembler post-processor
9012 cpp Options to pass to the C preprocessor
9013 cc1 Options to pass to the C compiler
9014 cc1plus Options to pass to the C++ compiler
9015 endfile Object files to include at the end of the link
9016 link Options to pass to the linker
9017 lib Libraries to include on the command line to the linker
9018 libgcc Decides which GCC support library to pass to the linker
9019 linker Sets the name of the linker
9020 predefines Defines to be passed to the C preprocessor
9021 signed_char Defines to pass to CPP to say whether @code{char} is signed
9023 startfile Object files to include at the start of the link
9026 Here is a small example of a spec file:
9032 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
9035 This example renames the spec called @samp{lib} to @samp{old_lib} and
9036 then overrides the previous definition of @samp{lib} with a new one.
9037 The new definition adds in some extra command-line options before
9038 including the text of the old definition.
9040 @dfn{Spec strings} are a list of command-line options to be passed to their
9041 corresponding program. In addition, the spec strings can contain
9042 @samp{%}-prefixed sequences to substitute variable text or to
9043 conditionally insert text into the command line. Using these constructs
9044 it is possible to generate quite complex command lines.
9046 Here is a table of all defined @samp{%}-sequences for spec
9047 strings. Note that spaces are not generated automatically around the
9048 results of expanding these sequences. Therefore you can concatenate them
9049 together or combine them with constant text in a single argument.
9053 Substitute one @samp{%} into the program name or argument.
9056 Substitute the name of the input file being processed.
9059 Substitute the basename of the input file being processed.
9060 This is the substring up to (and not including) the last period
9061 and not including the directory.
9064 This is the same as @samp{%b}, but include the file suffix (text after
9068 Marks the argument containing or following the @samp{%d} as a
9069 temporary file name, so that that file will be deleted if GCC exits
9070 successfully. Unlike @samp{%g}, this contributes no text to the
9073 @item %g@var{suffix}
9074 Substitute a file name that has suffix @var{suffix} and is chosen
9075 once per compilation, and mark the argument in the same way as
9076 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
9077 name is now chosen in a way that is hard to predict even when previously
9078 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
9079 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
9080 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
9081 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
9082 was simply substituted with a file name chosen once per compilation,
9083 without regard to any appended suffix (which was therefore treated
9084 just like ordinary text), making such attacks more likely to succeed.
9086 @item %u@var{suffix}
9087 Like @samp{%g}, but generates a new temporary file name even if
9088 @samp{%u@var{suffix}} was already seen.
9090 @item %U@var{suffix}
9091 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
9092 new one if there is no such last file name. In the absence of any
9093 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
9094 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
9095 would involve the generation of two distinct file names, one
9096 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
9097 simply substituted with a file name chosen for the previous @samp{%u},
9098 without regard to any appended suffix.
9100 @item %j@var{suffix}
9101 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
9102 writable, and if save-temps is off; otherwise, substitute the name
9103 of a temporary file, just like @samp{%u}. This temporary file is not
9104 meant for communication between processes, but rather as a junk
9107 @item %|@var{suffix}
9108 @itemx %m@var{suffix}
9109 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
9110 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
9111 all. These are the two most common ways to instruct a program that it
9112 should read from standard input or write to standard output. If you
9113 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
9114 construct: see for example @file{f/lang-specs.h}.
9116 @item %.@var{SUFFIX}
9117 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
9118 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
9119 terminated by the next space or %.
9122 Marks the argument containing or following the @samp{%w} as the
9123 designated output file of this compilation. This puts the argument
9124 into the sequence of arguments that @samp{%o} will substitute later.
9127 Substitutes the names of all the output files, with spaces
9128 automatically placed around them. You should write spaces
9129 around the @samp{%o} as well or the results are undefined.
9130 @samp{%o} is for use in the specs for running the linker.
9131 Input files whose names have no recognized suffix are not compiled
9132 at all, but they are included among the output files, so they will
9136 Substitutes the suffix for object files. Note that this is
9137 handled specially when it immediately follows @samp{%g, %u, or %U},
9138 because of the need for those to form complete file names. The
9139 handling is such that @samp{%O} is treated exactly as if it had already
9140 been substituted, except that @samp{%g, %u, and %U} do not currently
9141 support additional @var{suffix} characters following @samp{%O} as they would
9142 following, for example, @samp{.o}.
9145 Substitutes the standard macro predefinitions for the
9146 current target machine. Use this when running @code{cpp}.
9149 Like @samp{%p}, but puts @samp{__} before and after the name of each
9150 predefined macro, except for macros that start with @samp{__} or with
9151 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
9155 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
9156 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
9157 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
9158 and @option{-imultilib} as necessary.
9161 Current argument is the name of a library or startup file of some sort.
9162 Search for that file in a standard list of directories and substitute
9163 the full name found. The current working directory is included in the
9164 list of directories scanned.
9167 Current argument is the name of a linker script. Search for that file
9168 in the current list of directories to scan for libraries. If the file
9169 is located insert a @option{--script} option into the command line
9170 followed by the full path name found. If the file is not found then
9171 generate an error message. Note: the current working directory is not
9175 Print @var{str} as an error message. @var{str} is terminated by a newline.
9176 Use this when inconsistent options are detected.
9179 Substitute the contents of spec string @var{name} at this point.
9182 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
9184 @item %x@{@var{option}@}
9185 Accumulate an option for @samp{%X}.
9188 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
9192 Output the accumulated assembler options specified by @option{-Wa}.
9195 Output the accumulated preprocessor options specified by @option{-Wp}.
9198 Process the @code{asm} spec. This is used to compute the
9199 switches to be passed to the assembler.
9202 Process the @code{asm_final} spec. This is a spec string for
9203 passing switches to an assembler post-processor, if such a program is
9207 Process the @code{link} spec. This is the spec for computing the
9208 command line passed to the linker. Typically it will make use of the
9209 @samp{%L %G %S %D and %E} sequences.
9212 Dump out a @option{-L} option for each directory that GCC believes might
9213 contain startup files. If the target supports multilibs then the
9214 current multilib directory will be prepended to each of these paths.
9217 Process the @code{lib} spec. This is a spec string for deciding which
9218 libraries should be included on the command line to the linker.
9221 Process the @code{libgcc} spec. This is a spec string for deciding
9222 which GCC support library should be included on the command line to the linker.
9225 Process the @code{startfile} spec. This is a spec for deciding which
9226 object files should be the first ones passed to the linker. Typically
9227 this might be a file named @file{crt0.o}.
9230 Process the @code{endfile} spec. This is a spec string that specifies
9231 the last object files that will be passed to the linker.
9234 Process the @code{cpp} spec. This is used to construct the arguments
9235 to be passed to the C preprocessor.
9238 Process the @code{cc1} spec. This is used to construct the options to be
9239 passed to the actual C compiler (@samp{cc1}).
9242 Process the @code{cc1plus} spec. This is used to construct the options to be
9243 passed to the actual C++ compiler (@samp{cc1plus}).
9246 Substitute the variable part of a matched option. See below.
9247 Note that each comma in the substituted string is replaced by
9251 Remove all occurrences of @code{-S} from the command line. Note---this
9252 command is position dependent. @samp{%} commands in the spec string
9253 before this one will see @code{-S}, @samp{%} commands in the spec string
9254 after this one will not.
9256 @item %:@var{function}(@var{args})
9257 Call the named function @var{function}, passing it @var{args}.
9258 @var{args} is first processed as a nested spec string, then split
9259 into an argument vector in the usual fashion. The function returns
9260 a string which is processed as if it had appeared literally as part
9261 of the current spec.
9263 The following built-in spec functions are provided:
9267 The @code{getenv} spec function takes two arguments: an environment
9268 variable name and a string. If the environment variable is not
9269 defined, a fatal error is issued. Otherwise, the return value is the
9270 value of the environment variable concatenated with the string. For
9271 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
9274 %:getenv(TOPDIR /include)
9277 expands to @file{/path/to/top/include}.
9279 @item @code{if-exists}
9280 The @code{if-exists} spec function takes one argument, an absolute
9281 pathname to a file. If the file exists, @code{if-exists} returns the
9282 pathname. Here is a small example of its usage:
9286 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
9289 @item @code{if-exists-else}
9290 The @code{if-exists-else} spec function is similar to the @code{if-exists}
9291 spec function, except that it takes two arguments. The first argument is
9292 an absolute pathname to a file. If the file exists, @code{if-exists-else}
9293 returns the pathname. If it does not exist, it returns the second argument.
9294 This way, @code{if-exists-else} can be used to select one file or another,
9295 based on the existence of the first. Here is a small example of its usage:
9299 crt0%O%s %:if-exists(crti%O%s) \
9300 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
9303 @item @code{replace-outfile}
9304 The @code{replace-outfile} spec function takes two arguments. It looks for the
9305 first argument in the outfiles array and replaces it with the second argument. Here
9306 is a small example of its usage:
9309 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
9312 @item @code{print-asm-header}
9313 The @code{print-asm-header} function takes no arguments and simply
9314 prints a banner like:
9320 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
9323 It is used to separate compiler options from assembler options
9324 in the @option{--target-help} output.
9328 Substitutes the @code{-S} switch, if that switch was given to GCC@.
9329 If that switch was not specified, this substitutes nothing. Note that
9330 the leading dash is omitted when specifying this option, and it is
9331 automatically inserted if the substitution is performed. Thus the spec
9332 string @samp{%@{foo@}} would match the command-line option @option{-foo}
9333 and would output the command line option @option{-foo}.
9335 @item %W@{@code{S}@}
9336 Like %@{@code{S}@} but mark last argument supplied within as a file to be
9339 @item %@{@code{S}*@}
9340 Substitutes all the switches specified to GCC whose names start
9341 with @code{-S}, but which also take an argument. This is used for
9342 switches like @option{-o}, @option{-D}, @option{-I}, etc.
9343 GCC considers @option{-o foo} as being
9344 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
9345 text, including the space. Thus two arguments would be generated.
9347 @item %@{@code{S}*&@code{T}*@}
9348 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
9349 (the order of @code{S} and @code{T} in the spec is not significant).
9350 There can be any number of ampersand-separated variables; for each the
9351 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
9353 @item %@{@code{S}:@code{X}@}
9354 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
9356 @item %@{!@code{S}:@code{X}@}
9357 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
9359 @item %@{@code{S}*:@code{X}@}
9360 Substitutes @code{X} if one or more switches whose names start with
9361 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
9362 once, no matter how many such switches appeared. However, if @code{%*}
9363 appears somewhere in @code{X}, then @code{X} will be substituted once
9364 for each matching switch, with the @code{%*} replaced by the part of
9365 that switch that matched the @code{*}.
9367 @item %@{.@code{S}:@code{X}@}
9368 Substitutes @code{X}, if processing a file with suffix @code{S}.
9370 @item %@{!.@code{S}:@code{X}@}
9371 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
9373 @item %@{,@code{S}:@code{X}@}
9374 Substitutes @code{X}, if processing a file for language @code{S}.
9376 @item %@{!,@code{S}:@code{X}@}
9377 Substitutes @code{X}, if not processing a file for language @code{S}.
9379 @item %@{@code{S}|@code{P}:@code{X}@}
9380 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
9381 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
9382 @code{*} sequences as well, although they have a stronger binding than
9383 the @samp{|}. If @code{%*} appears in @code{X}, all of the
9384 alternatives must be starred, and only the first matching alternative
9387 For example, a spec string like this:
9390 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
9393 will output the following command-line options from the following input
9394 command-line options:
9399 -d fred.c -foo -baz -boggle
9400 -d jim.d -bar -baz -boggle
9403 @item %@{S:X; T:Y; :D@}
9405 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
9406 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
9407 be as many clauses as you need. This may be combined with @code{.},
9408 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
9413 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
9414 construct may contain other nested @samp{%} constructs or spaces, or
9415 even newlines. They are processed as usual, as described above.
9416 Trailing white space in @code{X} is ignored. White space may also
9417 appear anywhere on the left side of the colon in these constructs,
9418 except between @code{.} or @code{*} and the corresponding word.
9420 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
9421 handled specifically in these constructs. If another value of
9422 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
9423 @option{-W} switch is found later in the command line, the earlier
9424 switch value is ignored, except with @{@code{S}*@} where @code{S} is
9425 just one letter, which passes all matching options.
9427 The character @samp{|} at the beginning of the predicate text is used to
9428 indicate that a command should be piped to the following command, but
9429 only if @option{-pipe} is specified.
9431 It is built into GCC which switches take arguments and which do not.
9432 (You might think it would be useful to generalize this to allow each
9433 compiler's spec to say which switches take arguments. But this cannot
9434 be done in a consistent fashion. GCC cannot even decide which input
9435 files have been specified without knowing which switches take arguments,
9436 and it must know which input files to compile in order to tell which
9439 GCC also knows implicitly that arguments starting in @option{-l} are to be
9440 treated as compiler output files, and passed to the linker in their
9441 proper position among the other output files.
9443 @c man begin OPTIONS
9445 @node Target Options
9446 @section Specifying Target Machine and Compiler Version
9447 @cindex target options
9448 @cindex cross compiling
9449 @cindex specifying machine version
9450 @cindex specifying compiler version and target machine
9451 @cindex compiler version, specifying
9452 @cindex target machine, specifying
9454 The usual way to run GCC is to run the executable called @file{gcc}, or
9455 @file{<machine>-gcc} when cross-compiling, or
9456 @file{<machine>-gcc-<version>} to run a version other than the one that
9457 was installed last. Sometimes this is inconvenient, so GCC provides
9458 options that will switch to another cross-compiler or version.
9461 @item -b @var{machine}
9463 The argument @var{machine} specifies the target machine for compilation.
9465 The value to use for @var{machine} is the same as was specified as the
9466 machine type when configuring GCC as a cross-compiler. For
9467 example, if a cross-compiler was configured with @samp{configure
9468 arm-elf}, meaning to compile for an arm processor with elf binaries,
9469 then you would specify @option{-b arm-elf} to run that cross compiler.
9470 Because there are other options beginning with @option{-b}, the
9471 configuration must contain a hyphen, or @option{-b} alone should be one
9472 argument followed by the configuration in the next argument.
9474 @item -V @var{version}
9476 The argument @var{version} specifies which version of GCC to run.
9477 This is useful when multiple versions are installed. For example,
9478 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
9481 The @option{-V} and @option{-b} options work by running the
9482 @file{<machine>-gcc-<version>} executable, so there's no real reason to
9483 use them if you can just run that directly.
9485 @node Submodel Options
9486 @section Hardware Models and Configurations
9487 @cindex submodel options
9488 @cindex specifying hardware config
9489 @cindex hardware models and configurations, specifying
9490 @cindex machine dependent options
9492 Earlier we discussed the standard option @option{-b} which chooses among
9493 different installed compilers for completely different target
9494 machines, such as VAX vs.@: 68000 vs.@: 80386.
9496 In addition, each of these target machine types can have its own
9497 special options, starting with @samp{-m}, to choose among various
9498 hardware models or configurations---for example, 68010 vs 68020,
9499 floating coprocessor or none. A single installed version of the
9500 compiler can compile for any model or configuration, according to the
9503 Some configurations of the compiler also support additional special
9504 options, usually for compatibility with other compilers on the same
9507 @c This list is ordered alphanumerically by subsection name.
9508 @c It should be the same order and spelling as these options are listed
9509 @c in Machine Dependent Options
9515 * Blackfin Options::
9519 * DEC Alpha Options::
9520 * DEC Alpha/VMS Options::
9523 * GNU/Linux Options::
9526 * i386 and x86-64 Options::
9527 * i386 and x86-64 Windows Options::
9529 * IA-64/VMS Options::
9540 * picoChip Options::
9542 * RS/6000 and PowerPC Options::
9544 * S/390 and zSeries Options::
9549 * System V Options::
9554 * Xstormy16 Options::
9560 @subsection ARC Options
9563 These options are defined for ARC implementations:
9568 Compile code for little endian mode. This is the default.
9572 Compile code for big endian mode.
9575 @opindex mmangle-cpu
9576 Prepend the name of the cpu to all public symbol names.
9577 In multiple-processor systems, there are many ARC variants with different
9578 instruction and register set characteristics. This flag prevents code
9579 compiled for one cpu to be linked with code compiled for another.
9580 No facility exists for handling variants that are ``almost identical''.
9581 This is an all or nothing option.
9583 @item -mcpu=@var{cpu}
9585 Compile code for ARC variant @var{cpu}.
9586 Which variants are supported depend on the configuration.
9587 All variants support @option{-mcpu=base}, this is the default.
9589 @item -mtext=@var{text-section}
9590 @itemx -mdata=@var{data-section}
9591 @itemx -mrodata=@var{readonly-data-section}
9595 Put functions, data, and readonly data in @var{text-section},
9596 @var{data-section}, and @var{readonly-data-section} respectively
9597 by default. This can be overridden with the @code{section} attribute.
9598 @xref{Variable Attributes}.
9600 @item -mfix-cortex-m3-ldrd
9601 @opindex mfix-cortex-m3-ldrd
9602 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
9603 with overlapping destination and base registers are used. This option avoids
9604 generating these instructions. This option is enabled by default when
9605 @option{-mcpu=cortex-m3} is specified.
9610 @subsection ARM Options
9613 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
9617 @item -mabi=@var{name}
9619 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
9620 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
9623 @opindex mapcs-frame
9624 Generate a stack frame that is compliant with the ARM Procedure Call
9625 Standard for all functions, even if this is not strictly necessary for
9626 correct execution of the code. Specifying @option{-fomit-frame-pointer}
9627 with this option will cause the stack frames not to be generated for
9628 leaf functions. The default is @option{-mno-apcs-frame}.
9632 This is a synonym for @option{-mapcs-frame}.
9635 @c not currently implemented
9636 @item -mapcs-stack-check
9637 @opindex mapcs-stack-check
9638 Generate code to check the amount of stack space available upon entry to
9639 every function (that actually uses some stack space). If there is
9640 insufficient space available then either the function
9641 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
9642 called, depending upon the amount of stack space required. The run time
9643 system is required to provide these functions. The default is
9644 @option{-mno-apcs-stack-check}, since this produces smaller code.
9646 @c not currently implemented
9648 @opindex mapcs-float
9649 Pass floating point arguments using the float point registers. This is
9650 one of the variants of the APCS@. This option is recommended if the
9651 target hardware has a floating point unit or if a lot of floating point
9652 arithmetic is going to be performed by the code. The default is
9653 @option{-mno-apcs-float}, since integer only code is slightly increased in
9654 size if @option{-mapcs-float} is used.
9656 @c not currently implemented
9657 @item -mapcs-reentrant
9658 @opindex mapcs-reentrant
9659 Generate reentrant, position independent code. The default is
9660 @option{-mno-apcs-reentrant}.
9663 @item -mthumb-interwork
9664 @opindex mthumb-interwork
9665 Generate code which supports calling between the ARM and Thumb
9666 instruction sets. Without this option the two instruction sets cannot
9667 be reliably used inside one program. The default is
9668 @option{-mno-thumb-interwork}, since slightly larger code is generated
9669 when @option{-mthumb-interwork} is specified.
9671 @item -mno-sched-prolog
9672 @opindex mno-sched-prolog
9673 Prevent the reordering of instructions in the function prolog, or the
9674 merging of those instruction with the instructions in the function's
9675 body. This means that all functions will start with a recognizable set
9676 of instructions (or in fact one of a choice from a small set of
9677 different function prologues), and this information can be used to
9678 locate the start if functions inside an executable piece of code. The
9679 default is @option{-msched-prolog}.
9681 @item -mfloat-abi=@var{name}
9683 Specifies which floating-point ABI to use. Permissible values
9684 are: @samp{soft}, @samp{softfp} and @samp{hard}.
9686 Specifying @samp{soft} causes GCC to generate output containing
9687 library calls for floating-point operations.
9688 @samp{softfp} allows the generation of code using hardware floating-point
9689 instructions, but still uses the soft-float calling conventions.
9690 @samp{hard} allows generation of floating-point instructions
9691 and uses FPU-specific calling conventions.
9693 The default depends on the specific target configuration. Note that
9694 the hard-float and soft-float ABIs are not link-compatible; you must
9695 compile your entire program with the same ABI, and link with a
9696 compatible set of libraries.
9699 @opindex mhard-float
9700 Equivalent to @option{-mfloat-abi=hard}.
9703 @opindex msoft-float
9704 Equivalent to @option{-mfloat-abi=soft}.
9706 @item -mlittle-endian
9707 @opindex mlittle-endian
9708 Generate code for a processor running in little-endian mode. This is
9709 the default for all standard configurations.
9712 @opindex mbig-endian
9713 Generate code for a processor running in big-endian mode; the default is
9714 to compile code for a little-endian processor.
9716 @item -mwords-little-endian
9717 @opindex mwords-little-endian
9718 This option only applies when generating code for big-endian processors.
9719 Generate code for a little-endian word order but a big-endian byte
9720 order. That is, a byte order of the form @samp{32107654}. Note: this
9721 option should only be used if you require compatibility with code for
9722 big-endian ARM processors generated by versions of the compiler prior to
9725 @item -mcpu=@var{name}
9727 This specifies the name of the target ARM processor. GCC uses this name
9728 to determine what kind of instructions it can emit when generating
9729 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
9730 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
9731 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
9732 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
9733 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
9735 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
9736 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
9737 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
9738 @samp{strongarm1110},
9739 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
9740 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
9741 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
9742 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
9743 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
9744 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
9745 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
9746 @samp{cortex-a8}, @samp{cortex-a9},
9747 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-m3},
9750 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9752 @item -mtune=@var{name}
9754 This option is very similar to the @option{-mcpu=} option, except that
9755 instead of specifying the actual target processor type, and hence
9756 restricting which instructions can be used, it specifies that GCC should
9757 tune the performance of the code as if the target were of the type
9758 specified in this option, but still choosing the instructions that it
9759 will generate based on the cpu specified by a @option{-mcpu=} option.
9760 For some ARM implementations better performance can be obtained by using
9763 @item -march=@var{name}
9765 This specifies the name of the target ARM architecture. GCC uses this
9766 name to determine what kind of instructions it can emit when generating
9767 assembly code. This option can be used in conjunction with or instead
9768 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
9769 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
9770 @samp{armv5}, @samp{armv5t}, @samp{armv5e}, @samp{armv5te},
9771 @samp{armv6}, @samp{armv6j},
9772 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
9773 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
9774 @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9776 @item -mfpu=@var{name}
9777 @itemx -mfpe=@var{number}
9778 @itemx -mfp=@var{number}
9782 This specifies what floating point hardware (or hardware emulation) is
9783 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
9784 @samp{fpe3}, @samp{maverick}, @samp{vfp}, @samp{vfpv3}, @samp{vfpv3-d16},
9785 @samp{neon}, and @samp{neon-fp16}. @option{-mfp} and @option{-mfpe}
9786 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
9787 with older versions of GCC@.
9789 If @option{-msoft-float} is specified this specifies the format of
9790 floating point values.
9792 @item -mfp16-format=@var{name}
9793 @opindex mfp16-format
9794 Specify the format of the @code{__fp16} half-precision floating-point type.
9795 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
9796 the default is @samp{none}, in which case the @code{__fp16} type is not
9797 defined. @xref{Half-Precision}, for more information.
9799 @item -mstructure-size-boundary=@var{n}
9800 @opindex mstructure-size-boundary
9801 The size of all structures and unions will be rounded up to a multiple
9802 of the number of bits set by this option. Permissible values are 8, 32
9803 and 64. The default value varies for different toolchains. For the COFF
9804 targeted toolchain the default value is 8. A value of 64 is only allowed
9805 if the underlying ABI supports it.
9807 Specifying the larger number can produce faster, more efficient code, but
9808 can also increase the size of the program. Different values are potentially
9809 incompatible. Code compiled with one value cannot necessarily expect to
9810 work with code or libraries compiled with another value, if they exchange
9811 information using structures or unions.
9813 @item -mabort-on-noreturn
9814 @opindex mabort-on-noreturn
9815 Generate a call to the function @code{abort} at the end of a
9816 @code{noreturn} function. It will be executed if the function tries to
9820 @itemx -mno-long-calls
9821 @opindex mlong-calls
9822 @opindex mno-long-calls
9823 Tells the compiler to perform function calls by first loading the
9824 address of the function into a register and then performing a subroutine
9825 call on this register. This switch is needed if the target function
9826 will lie outside of the 64 megabyte addressing range of the offset based
9827 version of subroutine call instruction.
9829 Even if this switch is enabled, not all function calls will be turned
9830 into long calls. The heuristic is that static functions, functions
9831 which have the @samp{short-call} attribute, functions that are inside
9832 the scope of a @samp{#pragma no_long_calls} directive and functions whose
9833 definitions have already been compiled within the current compilation
9834 unit, will not be turned into long calls. The exception to this rule is
9835 that weak function definitions, functions with the @samp{long-call}
9836 attribute or the @samp{section} attribute, and functions that are within
9837 the scope of a @samp{#pragma long_calls} directive, will always be
9838 turned into long calls.
9840 This feature is not enabled by default. Specifying
9841 @option{-mno-long-calls} will restore the default behavior, as will
9842 placing the function calls within the scope of a @samp{#pragma
9843 long_calls_off} directive. Note these switches have no effect on how
9844 the compiler generates code to handle function calls via function
9847 @item -msingle-pic-base
9848 @opindex msingle-pic-base
9849 Treat the register used for PIC addressing as read-only, rather than
9850 loading it in the prologue for each function. The run-time system is
9851 responsible for initializing this register with an appropriate value
9852 before execution begins.
9854 @item -mpic-register=@var{reg}
9855 @opindex mpic-register
9856 Specify the register to be used for PIC addressing. The default is R10
9857 unless stack-checking is enabled, when R9 is used.
9859 @item -mcirrus-fix-invalid-insns
9860 @opindex mcirrus-fix-invalid-insns
9861 @opindex mno-cirrus-fix-invalid-insns
9862 Insert NOPs into the instruction stream to in order to work around
9863 problems with invalid Maverick instruction combinations. This option
9864 is only valid if the @option{-mcpu=ep9312} option has been used to
9865 enable generation of instructions for the Cirrus Maverick floating
9866 point co-processor. This option is not enabled by default, since the
9867 problem is only present in older Maverick implementations. The default
9868 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
9871 @item -mpoke-function-name
9872 @opindex mpoke-function-name
9873 Write the name of each function into the text section, directly
9874 preceding the function prologue. The generated code is similar to this:
9878 .ascii "arm_poke_function_name", 0
9881 .word 0xff000000 + (t1 - t0)
9882 arm_poke_function_name
9884 stmfd sp!, @{fp, ip, lr, pc@}
9888 When performing a stack backtrace, code can inspect the value of
9889 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
9890 location @code{pc - 12} and the top 8 bits are set, then we know that
9891 there is a function name embedded immediately preceding this location
9892 and has length @code{((pc[-3]) & 0xff000000)}.
9896 Generate code for the Thumb instruction set. The default is to
9897 use the 32-bit ARM instruction set.
9898 This option automatically enables either 16-bit Thumb-1 or
9899 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
9900 and @option{-march=@var{name}} options. This option is not passed to the
9901 assembler. If you want to force assembler files to be interpreted as Thumb code,
9902 either add a @samp{.thumb} directive to the source or pass the @option{-mthumb}
9903 option directly to the assembler by prefixing it with @option{-Wa}.
9906 @opindex mtpcs-frame
9907 Generate a stack frame that is compliant with the Thumb Procedure Call
9908 Standard for all non-leaf functions. (A leaf function is one that does
9909 not call any other functions.) The default is @option{-mno-tpcs-frame}.
9911 @item -mtpcs-leaf-frame
9912 @opindex mtpcs-leaf-frame
9913 Generate a stack frame that is compliant with the Thumb Procedure Call
9914 Standard for all leaf functions. (A leaf function is one that does
9915 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
9917 @item -mcallee-super-interworking
9918 @opindex mcallee-super-interworking
9919 Gives all externally visible functions in the file being compiled an ARM
9920 instruction set header which switches to Thumb mode before executing the
9921 rest of the function. This allows these functions to be called from
9922 non-interworking code. This option is not valid in AAPCS configurations
9923 because interworking is enabled by default.
9925 @item -mcaller-super-interworking
9926 @opindex mcaller-super-interworking
9927 Allows calls via function pointers (including virtual functions) to
9928 execute correctly regardless of whether the target code has been
9929 compiled for interworking or not. There is a small overhead in the cost
9930 of executing a function pointer if this option is enabled. This option
9931 is not valid in AAPCS configurations because interworking is enabled
9934 @item -mtp=@var{name}
9936 Specify the access model for the thread local storage pointer. The valid
9937 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
9938 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
9939 (supported in the arm6k architecture), and @option{auto}, which uses the
9940 best available method for the selected processor. The default setting is
9943 @item -mword-relocations
9944 @opindex mword-relocations
9945 Only generate absolute relocations on word sized values (i.e. R_ARM_ABS32).
9946 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
9947 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
9953 @subsection AVR Options
9956 These options are defined for AVR implementations:
9959 @item -mmcu=@var{mcu}
9961 Specify ATMEL AVR instruction set or MCU type.
9963 Instruction set avr1 is for the minimal AVR core, not supported by the C
9964 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9965 attiny11, attiny12, attiny15, attiny28).
9967 Instruction set avr2 (default) is for the classic AVR core with up to
9968 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9969 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9970 at90c8534, at90s8535).
9972 Instruction set avr3 is for the classic AVR core with up to 128K program
9973 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9975 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9976 memory space (MCU types: atmega8, atmega83, atmega85).
9978 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9979 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9980 atmega64, atmega128, at43usb355, at94k).
9982 @item -mno-interrupts
9983 @opindex mno-interrupts
9984 Generated code is not compatible with hardware interrupts.
9985 Code size will be smaller.
9987 @item -mcall-prologues
9988 @opindex mcall-prologues
9989 Functions prologues/epilogues expanded as call to appropriate
9990 subroutines. Code size will be smaller.
9993 @opindex mtiny-stack
9994 Change only the low 8 bits of the stack pointer.
9998 Assume int to be 8 bit integer. This affects the sizes of all types: A
9999 char will be 1 byte, an int will be 1 byte, a long will be 2 bytes
10000 and long long will be 4 bytes. Please note that this option does not
10001 comply to the C standards, but it will provide you with smaller code
10005 @node Blackfin Options
10006 @subsection Blackfin Options
10007 @cindex Blackfin Options
10010 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
10012 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
10013 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
10014 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
10015 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
10016 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
10017 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
10018 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
10020 The optional @var{sirevision} specifies the silicon revision of the target
10021 Blackfin processor. Any workarounds available for the targeted silicon revision
10022 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
10023 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
10024 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
10025 hexadecimal digits representing the major and minor numbers in the silicon
10026 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
10027 is not defined. If @var{sirevision} is @samp{any}, the
10028 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
10029 If this optional @var{sirevision} is not used, GCC assumes the latest known
10030 silicon revision of the targeted Blackfin processor.
10032 Support for @samp{bf561} is incomplete. For @samp{bf561},
10033 Only the processor macro is defined.
10034 Without this option, @samp{bf532} is used as the processor by default.
10035 The corresponding predefined processor macros for @var{cpu} is to
10036 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
10037 provided by libgloss to be linked in if @option{-msim} is not given.
10041 Specifies that the program will be run on the simulator. This causes
10042 the simulator BSP provided by libgloss to be linked in. This option
10043 has effect only for @samp{bfin-elf} toolchain.
10044 Certain other options, such as @option{-mid-shared-library} and
10045 @option{-mfdpic}, imply @option{-msim}.
10047 @item -momit-leaf-frame-pointer
10048 @opindex momit-leaf-frame-pointer
10049 Don't keep the frame pointer in a register for leaf functions. This
10050 avoids the instructions to save, set up and restore frame pointers and
10051 makes an extra register available in leaf functions. The option
10052 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10053 which might make debugging harder.
10055 @item -mspecld-anomaly
10056 @opindex mspecld-anomaly
10057 When enabled, the compiler will ensure that the generated code does not
10058 contain speculative loads after jump instructions. If this option is used,
10059 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
10061 @item -mno-specld-anomaly
10062 @opindex mno-specld-anomaly
10063 Don't generate extra code to prevent speculative loads from occurring.
10065 @item -mcsync-anomaly
10066 @opindex mcsync-anomaly
10067 When enabled, the compiler will ensure that the generated code does not
10068 contain CSYNC or SSYNC instructions too soon after conditional branches.
10069 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
10071 @item -mno-csync-anomaly
10072 @opindex mno-csync-anomaly
10073 Don't generate extra code to prevent CSYNC or SSYNC instructions from
10074 occurring too soon after a conditional branch.
10078 When enabled, the compiler is free to take advantage of the knowledge that
10079 the entire program fits into the low 64k of memory.
10082 @opindex mno-low-64k
10083 Assume that the program is arbitrarily large. This is the default.
10085 @item -mstack-check-l1
10086 @opindex mstack-check-l1
10087 Do stack checking using information placed into L1 scratchpad memory by the
10090 @item -mid-shared-library
10091 @opindex mid-shared-library
10092 Generate code that supports shared libraries via the library ID method.
10093 This allows for execute in place and shared libraries in an environment
10094 without virtual memory management. This option implies @option{-fPIC}.
10095 With a @samp{bfin-elf} target, this option implies @option{-msim}.
10097 @item -mno-id-shared-library
10098 @opindex mno-id-shared-library
10099 Generate code that doesn't assume ID based shared libraries are being used.
10100 This is the default.
10102 @item -mleaf-id-shared-library
10103 @opindex mleaf-id-shared-library
10104 Generate code that supports shared libraries via the library ID method,
10105 but assumes that this library or executable won't link against any other
10106 ID shared libraries. That allows the compiler to use faster code for jumps
10109 @item -mno-leaf-id-shared-library
10110 @opindex mno-leaf-id-shared-library
10111 Do not assume that the code being compiled won't link against any ID shared
10112 libraries. Slower code will be generated for jump and call insns.
10114 @item -mshared-library-id=n
10115 @opindex mshared-library-id
10116 Specified the identification number of the ID based shared library being
10117 compiled. Specifying a value of 0 will generate more compact code, specifying
10118 other values will force the allocation of that number to the current
10119 library but is no more space or time efficient than omitting this option.
10123 Generate code that allows the data segment to be located in a different
10124 area of memory from the text segment. This allows for execute in place in
10125 an environment without virtual memory management by eliminating relocations
10126 against the text section.
10128 @item -mno-sep-data
10129 @opindex mno-sep-data
10130 Generate code that assumes that the data segment follows the text segment.
10131 This is the default.
10134 @itemx -mno-long-calls
10135 @opindex mlong-calls
10136 @opindex mno-long-calls
10137 Tells the compiler to perform function calls by first loading the
10138 address of the function into a register and then performing a subroutine
10139 call on this register. This switch is needed if the target function
10140 will lie outside of the 24 bit addressing range of the offset based
10141 version of subroutine call instruction.
10143 This feature is not enabled by default. Specifying
10144 @option{-mno-long-calls} will restore the default behavior. Note these
10145 switches have no effect on how the compiler generates code to handle
10146 function calls via function pointers.
10150 Link with the fast floating-point library. This library relaxes some of
10151 the IEEE floating-point standard's rules for checking inputs against
10152 Not-a-Number (NAN), in the interest of performance.
10155 @opindex minline-plt
10156 Enable inlining of PLT entries in function calls to functions that are
10157 not known to bind locally. It has no effect without @option{-mfdpic}.
10160 @opindex mmulticore
10161 Build standalone application for multicore Blackfin processor. Proper
10162 start files and link scripts will be used to support multicore.
10163 This option defines @code{__BFIN_MULTICORE}. It can only be used with
10164 @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}. It can be used with
10165 @option{-mcorea} or @option{-mcoreb}. If it's used without
10166 @option{-mcorea} or @option{-mcoreb}, single application/dual core
10167 programming model is used. In this model, the main function of Core B
10168 should be named as coreb_main. If it's used with @option{-mcorea} or
10169 @option{-mcoreb}, one application per core programming model is used.
10170 If this option is not used, single core application programming
10175 Build standalone application for Core A of BF561 when using
10176 one application per core programming model. Proper start files
10177 and link scripts will be used to support Core A. This option
10178 defines @code{__BFIN_COREA}. It must be used with @option{-mmulticore}.
10182 Build standalone application for Core B of BF561 when using
10183 one application per core programming model. Proper start files
10184 and link scripts will be used to support Core B. This option
10185 defines @code{__BFIN_COREB}. When this option is used, coreb_main
10186 should be used instead of main. It must be used with
10187 @option{-mmulticore}.
10191 Build standalone application for SDRAM. Proper start files and
10192 link scripts will be used to put the application into SDRAM.
10193 Loader should initialize SDRAM before loading the application
10194 into SDRAM. This option defines @code{__BFIN_SDRAM}.
10198 Assume that ICPLBs are enabled at runtime. This has an effect on certain
10199 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
10200 are enabled; for standalone applications the default is off.
10204 @subsection CRIS Options
10205 @cindex CRIS Options
10207 These options are defined specifically for the CRIS ports.
10210 @item -march=@var{architecture-type}
10211 @itemx -mcpu=@var{architecture-type}
10214 Generate code for the specified architecture. The choices for
10215 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10216 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
10217 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10220 @item -mtune=@var{architecture-type}
10222 Tune to @var{architecture-type} everything applicable about the generated
10223 code, except for the ABI and the set of available instructions. The
10224 choices for @var{architecture-type} are the same as for
10225 @option{-march=@var{architecture-type}}.
10227 @item -mmax-stack-frame=@var{n}
10228 @opindex mmax-stack-frame
10229 Warn when the stack frame of a function exceeds @var{n} bytes.
10235 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10236 @option{-march=v3} and @option{-march=v8} respectively.
10238 @item -mmul-bug-workaround
10239 @itemx -mno-mul-bug-workaround
10240 @opindex mmul-bug-workaround
10241 @opindex mno-mul-bug-workaround
10242 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
10243 models where it applies. This option is active by default.
10247 Enable CRIS-specific verbose debug-related information in the assembly
10248 code. This option also has the effect to turn off the @samp{#NO_APP}
10249 formatted-code indicator to the assembler at the beginning of the
10254 Do not use condition-code results from previous instruction; always emit
10255 compare and test instructions before use of condition codes.
10257 @item -mno-side-effects
10258 @opindex mno-side-effects
10259 Do not emit instructions with side-effects in addressing modes other than
10262 @item -mstack-align
10263 @itemx -mno-stack-align
10264 @itemx -mdata-align
10265 @itemx -mno-data-align
10266 @itemx -mconst-align
10267 @itemx -mno-const-align
10268 @opindex mstack-align
10269 @opindex mno-stack-align
10270 @opindex mdata-align
10271 @opindex mno-data-align
10272 @opindex mconst-align
10273 @opindex mno-const-align
10274 These options (no-options) arranges (eliminate arrangements) for the
10275 stack-frame, individual data and constants to be aligned for the maximum
10276 single data access size for the chosen CPU model. The default is to
10277 arrange for 32-bit alignment. ABI details such as structure layout are
10278 not affected by these options.
10286 Similar to the stack- data- and const-align options above, these options
10287 arrange for stack-frame, writable data and constants to all be 32-bit,
10288 16-bit or 8-bit aligned. The default is 32-bit alignment.
10290 @item -mno-prologue-epilogue
10291 @itemx -mprologue-epilogue
10292 @opindex mno-prologue-epilogue
10293 @opindex mprologue-epilogue
10294 With @option{-mno-prologue-epilogue}, the normal function prologue and
10295 epilogue that sets up the stack-frame are omitted and no return
10296 instructions or return sequences are generated in the code. Use this
10297 option only together with visual inspection of the compiled code: no
10298 warnings or errors are generated when call-saved registers must be saved,
10299 or storage for local variable needs to be allocated.
10303 @opindex mno-gotplt
10305 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10306 instruction sequences that load addresses for functions from the PLT part
10307 of the GOT rather than (traditional on other architectures) calls to the
10308 PLT@. The default is @option{-mgotplt}.
10312 Legacy no-op option only recognized with the cris-axis-elf and
10313 cris-axis-linux-gnu targets.
10317 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10321 This option, recognized for the cris-axis-elf arranges
10322 to link with input-output functions from a simulator library. Code,
10323 initialized data and zero-initialized data are allocated consecutively.
10327 Like @option{-sim}, but pass linker options to locate initialized data at
10328 0x40000000 and zero-initialized data at 0x80000000.
10332 @subsection CRX Options
10333 @cindex CRX Options
10335 These options are defined specifically for the CRX ports.
10341 Enable the use of multiply-accumulate instructions. Disabled by default.
10344 @opindex mpush-args
10345 Push instructions will be used to pass outgoing arguments when functions
10346 are called. Enabled by default.
10349 @node Darwin Options
10350 @subsection Darwin Options
10351 @cindex Darwin options
10353 These options are defined for all architectures running the Darwin operating
10356 FSF GCC on Darwin does not create ``fat'' object files; it will create
10357 an object file for the single architecture that it was built to
10358 target. Apple's GCC on Darwin does create ``fat'' files if multiple
10359 @option{-arch} options are used; it does so by running the compiler or
10360 linker multiple times and joining the results together with
10363 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
10364 @samp{i686}) is determined by the flags that specify the ISA
10365 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
10366 @option{-force_cpusubtype_ALL} option can be used to override this.
10368 The Darwin tools vary in their behavior when presented with an ISA
10369 mismatch. The assembler, @file{as}, will only permit instructions to
10370 be used that are valid for the subtype of the file it is generating,
10371 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
10372 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
10373 and print an error if asked to create a shared library with a less
10374 restrictive subtype than its input files (for instance, trying to put
10375 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
10376 for executables, @file{ld}, will quietly give the executable the most
10377 restrictive subtype of any of its input files.
10382 Add the framework directory @var{dir} to the head of the list of
10383 directories to be searched for header files. These directories are
10384 interleaved with those specified by @option{-I} options and are
10385 scanned in a left-to-right order.
10387 A framework directory is a directory with frameworks in it. A
10388 framework is a directory with a @samp{"Headers"} and/or
10389 @samp{"PrivateHeaders"} directory contained directly in it that ends
10390 in @samp{".framework"}. The name of a framework is the name of this
10391 directory excluding the @samp{".framework"}. Headers associated with
10392 the framework are found in one of those two directories, with
10393 @samp{"Headers"} being searched first. A subframework is a framework
10394 directory that is in a framework's @samp{"Frameworks"} directory.
10395 Includes of subframework headers can only appear in a header of a
10396 framework that contains the subframework, or in a sibling subframework
10397 header. Two subframeworks are siblings if they occur in the same
10398 framework. A subframework should not have the same name as a
10399 framework, a warning will be issued if this is violated. Currently a
10400 subframework cannot have subframeworks, in the future, the mechanism
10401 may be extended to support this. The standard frameworks can be found
10402 in @samp{"/System/Library/Frameworks"} and
10403 @samp{"/Library/Frameworks"}. An example include looks like
10404 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
10405 the name of the framework and header.h is found in the
10406 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
10408 @item -iframework@var{dir}
10409 @opindex iframework
10410 Like @option{-F} except the directory is a treated as a system
10411 directory. The main difference between this @option{-iframework} and
10412 @option{-F} is that with @option{-iframework} the compiler does not
10413 warn about constructs contained within header files found via
10414 @var{dir}. This option is valid only for the C family of languages.
10418 Emit debugging information for symbols that are used. For STABS
10419 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
10420 This is by default ON@.
10424 Emit debugging information for all symbols and types.
10426 @item -mmacosx-version-min=@var{version}
10427 The earliest version of MacOS X that this executable will run on
10428 is @var{version}. Typical values of @var{version} include @code{10.1},
10429 @code{10.2}, and @code{10.3.9}.
10431 If the compiler was built to use the system's headers by default,
10432 then the default for this option is the system version on which the
10433 compiler is running, otherwise the default is to make choices which
10434 are compatible with as many systems and code bases as possible.
10438 Enable kernel development mode. The @option{-mkernel} option sets
10439 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
10440 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
10441 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
10442 applicable. This mode also sets @option{-mno-altivec},
10443 @option{-msoft-float}, @option{-fno-builtin} and
10444 @option{-mlong-branch} for PowerPC targets.
10446 @item -mone-byte-bool
10447 @opindex mone-byte-bool
10448 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
10449 By default @samp{sizeof(bool)} is @samp{4} when compiling for
10450 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
10451 option has no effect on x86.
10453 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
10454 to generate code that is not binary compatible with code generated
10455 without that switch. Using this switch may require recompiling all
10456 other modules in a program, including system libraries. Use this
10457 switch to conform to a non-default data model.
10459 @item -mfix-and-continue
10460 @itemx -ffix-and-continue
10461 @itemx -findirect-data
10462 @opindex mfix-and-continue
10463 @opindex ffix-and-continue
10464 @opindex findirect-data
10465 Generate code suitable for fast turn around development. Needed to
10466 enable gdb to dynamically load @code{.o} files into already running
10467 programs. @option{-findirect-data} and @option{-ffix-and-continue}
10468 are provided for backwards compatibility.
10472 Loads all members of static archive libraries.
10473 See man ld(1) for more information.
10475 @item -arch_errors_fatal
10476 @opindex arch_errors_fatal
10477 Cause the errors having to do with files that have the wrong architecture
10480 @item -bind_at_load
10481 @opindex bind_at_load
10482 Causes the output file to be marked such that the dynamic linker will
10483 bind all undefined references when the file is loaded or launched.
10487 Produce a Mach-o bundle format file.
10488 See man ld(1) for more information.
10490 @item -bundle_loader @var{executable}
10491 @opindex bundle_loader
10492 This option specifies the @var{executable} that will be loading the build
10493 output file being linked. See man ld(1) for more information.
10496 @opindex dynamiclib
10497 When passed this option, GCC will produce a dynamic library instead of
10498 an executable when linking, using the Darwin @file{libtool} command.
10500 @item -force_cpusubtype_ALL
10501 @opindex force_cpusubtype_ALL
10502 This causes GCC's output file to have the @var{ALL} subtype, instead of
10503 one controlled by the @option{-mcpu} or @option{-march} option.
10505 @item -allowable_client @var{client_name}
10506 @itemx -client_name
10507 @itemx -compatibility_version
10508 @itemx -current_version
10510 @itemx -dependency-file
10512 @itemx -dylinker_install_name
10514 @itemx -exported_symbols_list
10516 @itemx -flat_namespace
10517 @itemx -force_flat_namespace
10518 @itemx -headerpad_max_install_names
10521 @itemx -install_name
10522 @itemx -keep_private_externs
10523 @itemx -multi_module
10524 @itemx -multiply_defined
10525 @itemx -multiply_defined_unused
10527 @itemx -no_dead_strip_inits_and_terms
10528 @itemx -nofixprebinding
10529 @itemx -nomultidefs
10531 @itemx -noseglinkedit
10532 @itemx -pagezero_size
10534 @itemx -prebind_all_twolevel_modules
10535 @itemx -private_bundle
10536 @itemx -read_only_relocs
10538 @itemx -sectobjectsymbols
10542 @itemx -sectobjectsymbols
10545 @itemx -segs_read_only_addr
10546 @itemx -segs_read_write_addr
10547 @itemx -seg_addr_table
10548 @itemx -seg_addr_table_filename
10549 @itemx -seglinkedit
10551 @itemx -segs_read_only_addr
10552 @itemx -segs_read_write_addr
10553 @itemx -single_module
10555 @itemx -sub_library
10556 @itemx -sub_umbrella
10557 @itemx -twolevel_namespace
10560 @itemx -unexported_symbols_list
10561 @itemx -weak_reference_mismatches
10562 @itemx -whatsloaded
10563 @opindex allowable_client
10564 @opindex client_name
10565 @opindex compatibility_version
10566 @opindex current_version
10567 @opindex dead_strip
10568 @opindex dependency-file
10569 @opindex dylib_file
10570 @opindex dylinker_install_name
10572 @opindex exported_symbols_list
10574 @opindex flat_namespace
10575 @opindex force_flat_namespace
10576 @opindex headerpad_max_install_names
10577 @opindex image_base
10579 @opindex install_name
10580 @opindex keep_private_externs
10581 @opindex multi_module
10582 @opindex multiply_defined
10583 @opindex multiply_defined_unused
10584 @opindex noall_load
10585 @opindex no_dead_strip_inits_and_terms
10586 @opindex nofixprebinding
10587 @opindex nomultidefs
10589 @opindex noseglinkedit
10590 @opindex pagezero_size
10592 @opindex prebind_all_twolevel_modules
10593 @opindex private_bundle
10594 @opindex read_only_relocs
10596 @opindex sectobjectsymbols
10599 @opindex sectcreate
10600 @opindex sectobjectsymbols
10603 @opindex segs_read_only_addr
10604 @opindex segs_read_write_addr
10605 @opindex seg_addr_table
10606 @opindex seg_addr_table_filename
10607 @opindex seglinkedit
10609 @opindex segs_read_only_addr
10610 @opindex segs_read_write_addr
10611 @opindex single_module
10613 @opindex sub_library
10614 @opindex sub_umbrella
10615 @opindex twolevel_namespace
10618 @opindex unexported_symbols_list
10619 @opindex weak_reference_mismatches
10620 @opindex whatsloaded
10621 These options are passed to the Darwin linker. The Darwin linker man page
10622 describes them in detail.
10625 @node DEC Alpha Options
10626 @subsection DEC Alpha Options
10628 These @samp{-m} options are defined for the DEC Alpha implementations:
10631 @item -mno-soft-float
10632 @itemx -msoft-float
10633 @opindex mno-soft-float
10634 @opindex msoft-float
10635 Use (do not use) the hardware floating-point instructions for
10636 floating-point operations. When @option{-msoft-float} is specified,
10637 functions in @file{libgcc.a} will be used to perform floating-point
10638 operations. Unless they are replaced by routines that emulate the
10639 floating-point operations, or compiled in such a way as to call such
10640 emulations routines, these routines will issue floating-point
10641 operations. If you are compiling for an Alpha without floating-point
10642 operations, you must ensure that the library is built so as not to call
10645 Note that Alpha implementations without floating-point operations are
10646 required to have floating-point registers.
10649 @itemx -mno-fp-regs
10651 @opindex mno-fp-regs
10652 Generate code that uses (does not use) the floating-point register set.
10653 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
10654 register set is not used, floating point operands are passed in integer
10655 registers as if they were integers and floating-point results are passed
10656 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
10657 so any function with a floating-point argument or return value called by code
10658 compiled with @option{-mno-fp-regs} must also be compiled with that
10661 A typical use of this option is building a kernel that does not use,
10662 and hence need not save and restore, any floating-point registers.
10666 The Alpha architecture implements floating-point hardware optimized for
10667 maximum performance. It is mostly compliant with the IEEE floating
10668 point standard. However, for full compliance, software assistance is
10669 required. This option generates code fully IEEE compliant code
10670 @emph{except} that the @var{inexact-flag} is not maintained (see below).
10671 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
10672 defined during compilation. The resulting code is less efficient but is
10673 able to correctly support denormalized numbers and exceptional IEEE
10674 values such as not-a-number and plus/minus infinity. Other Alpha
10675 compilers call this option @option{-ieee_with_no_inexact}.
10677 @item -mieee-with-inexact
10678 @opindex mieee-with-inexact
10679 This is like @option{-mieee} except the generated code also maintains
10680 the IEEE @var{inexact-flag}. Turning on this option causes the
10681 generated code to implement fully-compliant IEEE math. In addition to
10682 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
10683 macro. On some Alpha implementations the resulting code may execute
10684 significantly slower than the code generated by default. Since there is
10685 very little code that depends on the @var{inexact-flag}, you should
10686 normally not specify this option. Other Alpha compilers call this
10687 option @option{-ieee_with_inexact}.
10689 @item -mfp-trap-mode=@var{trap-mode}
10690 @opindex mfp-trap-mode
10691 This option controls what floating-point related traps are enabled.
10692 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
10693 The trap mode can be set to one of four values:
10697 This is the default (normal) setting. The only traps that are enabled
10698 are the ones that cannot be disabled in software (e.g., division by zero
10702 In addition to the traps enabled by @samp{n}, underflow traps are enabled
10706 Like @samp{u}, but the instructions are marked to be safe for software
10707 completion (see Alpha architecture manual for details).
10710 Like @samp{su}, but inexact traps are enabled as well.
10713 @item -mfp-rounding-mode=@var{rounding-mode}
10714 @opindex mfp-rounding-mode
10715 Selects the IEEE rounding mode. Other Alpha compilers call this option
10716 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
10721 Normal IEEE rounding mode. Floating point numbers are rounded towards
10722 the nearest machine number or towards the even machine number in case
10726 Round towards minus infinity.
10729 Chopped rounding mode. Floating point numbers are rounded towards zero.
10732 Dynamic rounding mode. A field in the floating point control register
10733 (@var{fpcr}, see Alpha architecture reference manual) controls the
10734 rounding mode in effect. The C library initializes this register for
10735 rounding towards plus infinity. Thus, unless your program modifies the
10736 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
10739 @item -mtrap-precision=@var{trap-precision}
10740 @opindex mtrap-precision
10741 In the Alpha architecture, floating point traps are imprecise. This
10742 means without software assistance it is impossible to recover from a
10743 floating trap and program execution normally needs to be terminated.
10744 GCC can generate code that can assist operating system trap handlers
10745 in determining the exact location that caused a floating point trap.
10746 Depending on the requirements of an application, different levels of
10747 precisions can be selected:
10751 Program precision. This option is the default and means a trap handler
10752 can only identify which program caused a floating point exception.
10755 Function precision. The trap handler can determine the function that
10756 caused a floating point exception.
10759 Instruction precision. The trap handler can determine the exact
10760 instruction that caused a floating point exception.
10763 Other Alpha compilers provide the equivalent options called
10764 @option{-scope_safe} and @option{-resumption_safe}.
10766 @item -mieee-conformant
10767 @opindex mieee-conformant
10768 This option marks the generated code as IEEE conformant. You must not
10769 use this option unless you also specify @option{-mtrap-precision=i} and either
10770 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
10771 is to emit the line @samp{.eflag 48} in the function prologue of the
10772 generated assembly file. Under DEC Unix, this has the effect that
10773 IEEE-conformant math library routines will be linked in.
10775 @item -mbuild-constants
10776 @opindex mbuild-constants
10777 Normally GCC examines a 32- or 64-bit integer constant to
10778 see if it can construct it from smaller constants in two or three
10779 instructions. If it cannot, it will output the constant as a literal and
10780 generate code to load it from the data segment at runtime.
10782 Use this option to require GCC to construct @emph{all} integer constants
10783 using code, even if it takes more instructions (the maximum is six).
10785 You would typically use this option to build a shared library dynamic
10786 loader. Itself a shared library, it must relocate itself in memory
10787 before it can find the variables and constants in its own data segment.
10793 Select whether to generate code to be assembled by the vendor-supplied
10794 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
10812 Indicate whether GCC should generate code to use the optional BWX,
10813 CIX, FIX and MAX instruction sets. The default is to use the instruction
10814 sets supported by the CPU type specified via @option{-mcpu=} option or that
10815 of the CPU on which GCC was built if none was specified.
10818 @itemx -mfloat-ieee
10819 @opindex mfloat-vax
10820 @opindex mfloat-ieee
10821 Generate code that uses (does not use) VAX F and G floating point
10822 arithmetic instead of IEEE single and double precision.
10824 @item -mexplicit-relocs
10825 @itemx -mno-explicit-relocs
10826 @opindex mexplicit-relocs
10827 @opindex mno-explicit-relocs
10828 Older Alpha assemblers provided no way to generate symbol relocations
10829 except via assembler macros. Use of these macros does not allow
10830 optimal instruction scheduling. GNU binutils as of version 2.12
10831 supports a new syntax that allows the compiler to explicitly mark
10832 which relocations should apply to which instructions. This option
10833 is mostly useful for debugging, as GCC detects the capabilities of
10834 the assembler when it is built and sets the default accordingly.
10837 @itemx -mlarge-data
10838 @opindex msmall-data
10839 @opindex mlarge-data
10840 When @option{-mexplicit-relocs} is in effect, static data is
10841 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
10842 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
10843 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
10844 16-bit relocations off of the @code{$gp} register. This limits the
10845 size of the small data area to 64KB, but allows the variables to be
10846 directly accessed via a single instruction.
10848 The default is @option{-mlarge-data}. With this option the data area
10849 is limited to just below 2GB@. Programs that require more than 2GB of
10850 data must use @code{malloc} or @code{mmap} to allocate the data in the
10851 heap instead of in the program's data segment.
10853 When generating code for shared libraries, @option{-fpic} implies
10854 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
10857 @itemx -mlarge-text
10858 @opindex msmall-text
10859 @opindex mlarge-text
10860 When @option{-msmall-text} is used, the compiler assumes that the
10861 code of the entire program (or shared library) fits in 4MB, and is
10862 thus reachable with a branch instruction. When @option{-msmall-data}
10863 is used, the compiler can assume that all local symbols share the
10864 same @code{$gp} value, and thus reduce the number of instructions
10865 required for a function call from 4 to 1.
10867 The default is @option{-mlarge-text}.
10869 @item -mcpu=@var{cpu_type}
10871 Set the instruction set and instruction scheduling parameters for
10872 machine type @var{cpu_type}. You can specify either the @samp{EV}
10873 style name or the corresponding chip number. GCC supports scheduling
10874 parameters for the EV4, EV5 and EV6 family of processors and will
10875 choose the default values for the instruction set from the processor
10876 you specify. If you do not specify a processor type, GCC will default
10877 to the processor on which the compiler was built.
10879 Supported values for @var{cpu_type} are
10885 Schedules as an EV4 and has no instruction set extensions.
10889 Schedules as an EV5 and has no instruction set extensions.
10893 Schedules as an EV5 and supports the BWX extension.
10898 Schedules as an EV5 and supports the BWX and MAX extensions.
10902 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
10906 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
10909 Native Linux/GNU toolchains also support the value @samp{native},
10910 which selects the best architecture option for the host processor.
10911 @option{-mcpu=native} has no effect if GCC does not recognize
10914 @item -mtune=@var{cpu_type}
10916 Set only the instruction scheduling parameters for machine type
10917 @var{cpu_type}. The instruction set is not changed.
10919 Native Linux/GNU toolchains also support the value @samp{native},
10920 which selects the best architecture option for the host processor.
10921 @option{-mtune=native} has no effect if GCC does not recognize
10924 @item -mmemory-latency=@var{time}
10925 @opindex mmemory-latency
10926 Sets the latency the scheduler should assume for typical memory
10927 references as seen by the application. This number is highly
10928 dependent on the memory access patterns used by the application
10929 and the size of the external cache on the machine.
10931 Valid options for @var{time} are
10935 A decimal number representing clock cycles.
10941 The compiler contains estimates of the number of clock cycles for
10942 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
10943 (also called Dcache, Scache, and Bcache), as well as to main memory.
10944 Note that L3 is only valid for EV5.
10949 @node DEC Alpha/VMS Options
10950 @subsection DEC Alpha/VMS Options
10952 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
10955 @item -mvms-return-codes
10956 @opindex mvms-return-codes
10957 Return VMS condition codes from main. The default is to return POSIX
10958 style condition (e.g.@: error) codes.
10960 @item -mdebug-main=@var{prefix}
10961 @opindex mdebug-main=@var{prefix}
10962 Flag the first routine whose name starts with @var{prefix} as the main
10963 routine for the debugger.
10967 Default to 64bit memory allocation routines.
10971 @subsection FR30 Options
10972 @cindex FR30 Options
10974 These options are defined specifically for the FR30 port.
10978 @item -msmall-model
10979 @opindex msmall-model
10980 Use the small address space model. This can produce smaller code, but
10981 it does assume that all symbolic values and addresses will fit into a
10986 Assume that run-time support has been provided and so there is no need
10987 to include the simulator library (@file{libsim.a}) on the linker
10993 @subsection FRV Options
10994 @cindex FRV Options
11000 Only use the first 32 general purpose registers.
11005 Use all 64 general purpose registers.
11010 Use only the first 32 floating point registers.
11015 Use all 64 floating point registers
11018 @opindex mhard-float
11020 Use hardware instructions for floating point operations.
11023 @opindex msoft-float
11025 Use library routines for floating point operations.
11030 Dynamically allocate condition code registers.
11035 Do not try to dynamically allocate condition code registers, only
11036 use @code{icc0} and @code{fcc0}.
11041 Change ABI to use double word insns.
11046 Do not use double word instructions.
11051 Use floating point double instructions.
11054 @opindex mno-double
11056 Do not use floating point double instructions.
11061 Use media instructions.
11066 Do not use media instructions.
11071 Use multiply and add/subtract instructions.
11074 @opindex mno-muladd
11076 Do not use multiply and add/subtract instructions.
11081 Select the FDPIC ABI, that uses function descriptors to represent
11082 pointers to functions. Without any PIC/PIE-related options, it
11083 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
11084 assumes GOT entries and small data are within a 12-bit range from the
11085 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
11086 are computed with 32 bits.
11087 With a @samp{bfin-elf} target, this option implies @option{-msim}.
11090 @opindex minline-plt
11092 Enable inlining of PLT entries in function calls to functions that are
11093 not known to bind locally. It has no effect without @option{-mfdpic}.
11094 It's enabled by default if optimizing for speed and compiling for
11095 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
11096 optimization option such as @option{-O3} or above is present in the
11102 Assume a large TLS segment when generating thread-local code.
11107 Do not assume a large TLS segment when generating thread-local code.
11112 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
11113 that is known to be in read-only sections. It's enabled by default,
11114 except for @option{-fpic} or @option{-fpie}: even though it may help
11115 make the global offset table smaller, it trades 1 instruction for 4.
11116 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
11117 one of which may be shared by multiple symbols, and it avoids the need
11118 for a GOT entry for the referenced symbol, so it's more likely to be a
11119 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
11121 @item -multilib-library-pic
11122 @opindex multilib-library-pic
11124 Link with the (library, not FD) pic libraries. It's implied by
11125 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
11126 @option{-fpic} without @option{-mfdpic}. You should never have to use
11130 @opindex mlinked-fp
11132 Follow the EABI requirement of always creating a frame pointer whenever
11133 a stack frame is allocated. This option is enabled by default and can
11134 be disabled with @option{-mno-linked-fp}.
11137 @opindex mlong-calls
11139 Use indirect addressing to call functions outside the current
11140 compilation unit. This allows the functions to be placed anywhere
11141 within the 32-bit address space.
11143 @item -malign-labels
11144 @opindex malign-labels
11146 Try to align labels to an 8-byte boundary by inserting nops into the
11147 previous packet. This option only has an effect when VLIW packing
11148 is enabled. It doesn't create new packets; it merely adds nops to
11151 @item -mlibrary-pic
11152 @opindex mlibrary-pic
11154 Generate position-independent EABI code.
11159 Use only the first four media accumulator registers.
11164 Use all eight media accumulator registers.
11169 Pack VLIW instructions.
11174 Do not pack VLIW instructions.
11177 @opindex mno-eflags
11179 Do not mark ABI switches in e_flags.
11182 @opindex mcond-move
11184 Enable the use of conditional-move instructions (default).
11186 This switch is mainly for debugging the compiler and will likely be removed
11187 in a future version.
11189 @item -mno-cond-move
11190 @opindex mno-cond-move
11192 Disable the use of conditional-move instructions.
11194 This switch is mainly for debugging the compiler and will likely be removed
11195 in a future version.
11200 Enable the use of conditional set instructions (default).
11202 This switch is mainly for debugging the compiler and will likely be removed
11203 in a future version.
11208 Disable the use of conditional set instructions.
11210 This switch is mainly for debugging the compiler and will likely be removed
11211 in a future version.
11214 @opindex mcond-exec
11216 Enable the use of conditional execution (default).
11218 This switch is mainly for debugging the compiler and will likely be removed
11219 in a future version.
11221 @item -mno-cond-exec
11222 @opindex mno-cond-exec
11224 Disable the use of conditional execution.
11226 This switch is mainly for debugging the compiler and will likely be removed
11227 in a future version.
11229 @item -mvliw-branch
11230 @opindex mvliw-branch
11232 Run a pass to pack branches into VLIW instructions (default).
11234 This switch is mainly for debugging the compiler and will likely be removed
11235 in a future version.
11237 @item -mno-vliw-branch
11238 @opindex mno-vliw-branch
11240 Do not run a pass to pack branches into VLIW instructions.
11242 This switch is mainly for debugging the compiler and will likely be removed
11243 in a future version.
11245 @item -mmulti-cond-exec
11246 @opindex mmulti-cond-exec
11248 Enable optimization of @code{&&} and @code{||} in conditional execution
11251 This switch is mainly for debugging the compiler and will likely be removed
11252 in a future version.
11254 @item -mno-multi-cond-exec
11255 @opindex mno-multi-cond-exec
11257 Disable optimization of @code{&&} and @code{||} in conditional execution.
11259 This switch is mainly for debugging the compiler and will likely be removed
11260 in a future version.
11262 @item -mnested-cond-exec
11263 @opindex mnested-cond-exec
11265 Enable nested conditional execution optimizations (default).
11267 This switch is mainly for debugging the compiler and will likely be removed
11268 in a future version.
11270 @item -mno-nested-cond-exec
11271 @opindex mno-nested-cond-exec
11273 Disable nested conditional execution optimizations.
11275 This switch is mainly for debugging the compiler and will likely be removed
11276 in a future version.
11278 @item -moptimize-membar
11279 @opindex moptimize-membar
11281 This switch removes redundant @code{membar} instructions from the
11282 compiler generated code. It is enabled by default.
11284 @item -mno-optimize-membar
11285 @opindex mno-optimize-membar
11287 This switch disables the automatic removal of redundant @code{membar}
11288 instructions from the generated code.
11290 @item -mtomcat-stats
11291 @opindex mtomcat-stats
11293 Cause gas to print out tomcat statistics.
11295 @item -mcpu=@var{cpu}
11298 Select the processor type for which to generate code. Possible values are
11299 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
11300 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
11304 @node GNU/Linux Options
11305 @subsection GNU/Linux Options
11307 These @samp{-m} options are defined for GNU/Linux targets:
11312 Use the GNU C library instead of uClibc. This is the default except
11313 on @samp{*-*-linux-*uclibc*} targets.
11317 Use uClibc instead of the GNU C library. This is the default on
11318 @samp{*-*-linux-*uclibc*} targets.
11321 @node H8/300 Options
11322 @subsection H8/300 Options
11324 These @samp{-m} options are defined for the H8/300 implementations:
11329 Shorten some address references at link time, when possible; uses the
11330 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
11331 ld, Using ld}, for a fuller description.
11335 Generate code for the H8/300H@.
11339 Generate code for the H8S@.
11343 Generate code for the H8S and H8/300H in the normal mode. This switch
11344 must be used either with @option{-mh} or @option{-ms}.
11348 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
11352 Make @code{int} data 32 bits by default.
11355 @opindex malign-300
11356 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
11357 The default for the H8/300H and H8S is to align longs and floats on 4
11359 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
11360 This option has no effect on the H8/300.
11364 @subsection HPPA Options
11365 @cindex HPPA Options
11367 These @samp{-m} options are defined for the HPPA family of computers:
11370 @item -march=@var{architecture-type}
11372 Generate code for the specified architecture. The choices for
11373 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
11374 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
11375 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
11376 architecture option for your machine. Code compiled for lower numbered
11377 architectures will run on higher numbered architectures, but not the
11380 @item -mpa-risc-1-0
11381 @itemx -mpa-risc-1-1
11382 @itemx -mpa-risc-2-0
11383 @opindex mpa-risc-1-0
11384 @opindex mpa-risc-1-1
11385 @opindex mpa-risc-2-0
11386 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
11389 @opindex mbig-switch
11390 Generate code suitable for big switch tables. Use this option only if
11391 the assembler/linker complain about out of range branches within a switch
11394 @item -mjump-in-delay
11395 @opindex mjump-in-delay
11396 Fill delay slots of function calls with unconditional jump instructions
11397 by modifying the return pointer for the function call to be the target
11398 of the conditional jump.
11400 @item -mdisable-fpregs
11401 @opindex mdisable-fpregs
11402 Prevent floating point registers from being used in any manner. This is
11403 necessary for compiling kernels which perform lazy context switching of
11404 floating point registers. If you use this option and attempt to perform
11405 floating point operations, the compiler will abort.
11407 @item -mdisable-indexing
11408 @opindex mdisable-indexing
11409 Prevent the compiler from using indexing address modes. This avoids some
11410 rather obscure problems when compiling MIG generated code under MACH@.
11412 @item -mno-space-regs
11413 @opindex mno-space-regs
11414 Generate code that assumes the target has no space registers. This allows
11415 GCC to generate faster indirect calls and use unscaled index address modes.
11417 Such code is suitable for level 0 PA systems and kernels.
11419 @item -mfast-indirect-calls
11420 @opindex mfast-indirect-calls
11421 Generate code that assumes calls never cross space boundaries. This
11422 allows GCC to emit code which performs faster indirect calls.
11424 This option will not work in the presence of shared libraries or nested
11427 @item -mfixed-range=@var{register-range}
11428 @opindex mfixed-range
11429 Generate code treating the given register range as fixed registers.
11430 A fixed register is one that the register allocator can not use. This is
11431 useful when compiling kernel code. A register range is specified as
11432 two registers separated by a dash. Multiple register ranges can be
11433 specified separated by a comma.
11435 @item -mlong-load-store
11436 @opindex mlong-load-store
11437 Generate 3-instruction load and store sequences as sometimes required by
11438 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
11441 @item -mportable-runtime
11442 @opindex mportable-runtime
11443 Use the portable calling conventions proposed by HP for ELF systems.
11447 Enable the use of assembler directives only GAS understands.
11449 @item -mschedule=@var{cpu-type}
11451 Schedule code according to the constraints for the machine type
11452 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
11453 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
11454 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
11455 proper scheduling option for your machine. The default scheduling is
11459 @opindex mlinker-opt
11460 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
11461 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
11462 linkers in which they give bogus error messages when linking some programs.
11465 @opindex msoft-float
11466 Generate output containing library calls for floating point.
11467 @strong{Warning:} the requisite libraries are not available for all HPPA
11468 targets. Normally the facilities of the machine's usual C compiler are
11469 used, but this cannot be done directly in cross-compilation. You must make
11470 your own arrangements to provide suitable library functions for
11473 @option{-msoft-float} changes the calling convention in the output file;
11474 therefore, it is only useful if you compile @emph{all} of a program with
11475 this option. In particular, you need to compile @file{libgcc.a}, the
11476 library that comes with GCC, with @option{-msoft-float} in order for
11481 Generate the predefine, @code{_SIO}, for server IO@. The default is
11482 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
11483 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
11484 options are available under HP-UX and HI-UX@.
11488 Use GNU ld specific options. This passes @option{-shared} to ld when
11489 building a shared library. It is the default when GCC is configured,
11490 explicitly or implicitly, with the GNU linker. This option does not
11491 have any affect on which ld is called, it only changes what parameters
11492 are passed to that ld. The ld that is called is determined by the
11493 @option{--with-ld} configure option, GCC's program search path, and
11494 finally by the user's @env{PATH}. The linker used by GCC can be printed
11495 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
11496 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11500 Use HP ld specific options. This passes @option{-b} to ld when building
11501 a shared library and passes @option{+Accept TypeMismatch} to ld on all
11502 links. It is the default when GCC is configured, explicitly or
11503 implicitly, with the HP linker. This option does not have any affect on
11504 which ld is called, it only changes what parameters are passed to that
11505 ld. The ld that is called is determined by the @option{--with-ld}
11506 configure option, GCC's program search path, and finally by the user's
11507 @env{PATH}. The linker used by GCC can be printed using @samp{which
11508 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
11509 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11512 @opindex mno-long-calls
11513 Generate code that uses long call sequences. This ensures that a call
11514 is always able to reach linker generated stubs. The default is to generate
11515 long calls only when the distance from the call site to the beginning
11516 of the function or translation unit, as the case may be, exceeds a
11517 predefined limit set by the branch type being used. The limits for
11518 normal calls are 7,600,000 and 240,000 bytes, respectively for the
11519 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
11522 Distances are measured from the beginning of functions when using the
11523 @option{-ffunction-sections} option, or when using the @option{-mgas}
11524 and @option{-mno-portable-runtime} options together under HP-UX with
11527 It is normally not desirable to use this option as it will degrade
11528 performance. However, it may be useful in large applications,
11529 particularly when partial linking is used to build the application.
11531 The types of long calls used depends on the capabilities of the
11532 assembler and linker, and the type of code being generated. The
11533 impact on systems that support long absolute calls, and long pic
11534 symbol-difference or pc-relative calls should be relatively small.
11535 However, an indirect call is used on 32-bit ELF systems in pic code
11536 and it is quite long.
11538 @item -munix=@var{unix-std}
11540 Generate compiler predefines and select a startfile for the specified
11541 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
11542 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
11543 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
11544 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
11545 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
11548 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
11549 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
11550 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
11551 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
11552 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
11553 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
11555 It is @emph{important} to note that this option changes the interfaces
11556 for various library routines. It also affects the operational behavior
11557 of the C library. Thus, @emph{extreme} care is needed in using this
11560 Library code that is intended to operate with more than one UNIX
11561 standard must test, set and restore the variable @var{__xpg4_extended_mask}
11562 as appropriate. Most GNU software doesn't provide this capability.
11566 Suppress the generation of link options to search libdld.sl when the
11567 @option{-static} option is specified on HP-UX 10 and later.
11571 The HP-UX implementation of setlocale in libc has a dependency on
11572 libdld.sl. There isn't an archive version of libdld.sl. Thus,
11573 when the @option{-static} option is specified, special link options
11574 are needed to resolve this dependency.
11576 On HP-UX 10 and later, the GCC driver adds the necessary options to
11577 link with libdld.sl when the @option{-static} option is specified.
11578 This causes the resulting binary to be dynamic. On the 64-bit port,
11579 the linkers generate dynamic binaries by default in any case. The
11580 @option{-nolibdld} option can be used to prevent the GCC driver from
11581 adding these link options.
11585 Add support for multithreading with the @dfn{dce thread} library
11586 under HP-UX@. This option sets flags for both the preprocessor and
11590 @node i386 and x86-64 Options
11591 @subsection Intel 386 and AMD x86-64 Options
11592 @cindex i386 Options
11593 @cindex x86-64 Options
11594 @cindex Intel 386 Options
11595 @cindex AMD x86-64 Options
11597 These @samp{-m} options are defined for the i386 and x86-64 family of
11601 @item -mtune=@var{cpu-type}
11603 Tune to @var{cpu-type} everything applicable about the generated code, except
11604 for the ABI and the set of available instructions. The choices for
11605 @var{cpu-type} are:
11608 Produce code optimized for the most common IA32/AMD64/EM64T processors.
11609 If you know the CPU on which your code will run, then you should use
11610 the corresponding @option{-mtune} option instead of
11611 @option{-mtune=generic}. But, if you do not know exactly what CPU users
11612 of your application will have, then you should use this option.
11614 As new processors are deployed in the marketplace, the behavior of this
11615 option will change. Therefore, if you upgrade to a newer version of
11616 GCC, the code generated option will change to reflect the processors
11617 that were most common when that version of GCC was released.
11619 There is no @option{-march=generic} option because @option{-march}
11620 indicates the instruction set the compiler can use, and there is no
11621 generic instruction set applicable to all processors. In contrast,
11622 @option{-mtune} indicates the processor (or, in this case, collection of
11623 processors) for which the code is optimized.
11625 This selects the CPU to tune for at compilation time by determining
11626 the processor type of the compiling machine. Using @option{-mtune=native}
11627 will produce code optimized for the local machine under the constraints
11628 of the selected instruction set. Using @option{-march=native} will
11629 enable all instruction subsets supported by the local machine (hence
11630 the result might not run on different machines).
11632 Original Intel's i386 CPU@.
11634 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
11635 @item i586, pentium
11636 Intel Pentium CPU with no MMX support.
11638 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
11640 Intel PentiumPro CPU@.
11642 Same as @code{generic}, but when used as @code{march} option, PentiumPro
11643 instruction set will be used, so the code will run on all i686 family chips.
11645 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
11646 @item pentium3, pentium3m
11647 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
11650 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
11651 support. Used by Centrino notebooks.
11652 @item pentium4, pentium4m
11653 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
11655 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
11658 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
11659 SSE2 and SSE3 instruction set support.
11661 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11662 instruction set support.
11664 Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11665 instruction set support.
11667 AMD K6 CPU with MMX instruction set support.
11669 Improved versions of AMD K6 CPU with MMX and 3dNOW!@: instruction set support.
11670 @item athlon, athlon-tbird
11671 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and SSE prefetch instructions
11673 @item athlon-4, athlon-xp, athlon-mp
11674 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and full SSE
11675 instruction set support.
11676 @item k8, opteron, athlon64, athlon-fx
11677 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
11678 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW!@: and 64-bit instruction set extensions.)
11679 @item k8-sse3, opteron-sse3, athlon64-sse3
11680 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
11681 @item amdfam10, barcelona
11682 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
11683 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
11684 instruction set extensions.)
11686 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
11689 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!@:
11690 instruction set support.
11692 Via C3 CPU with MMX and 3dNOW!@: instruction set support. (No scheduling is
11693 implemented for this chip.)
11695 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
11696 implemented for this chip.)
11698 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
11701 While picking a specific @var{cpu-type} will schedule things appropriately
11702 for that particular chip, the compiler will not generate any code that
11703 does not run on the i386 without the @option{-march=@var{cpu-type}} option
11706 @item -march=@var{cpu-type}
11708 Generate instructions for the machine type @var{cpu-type}. The choices
11709 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
11710 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
11712 @item -mcpu=@var{cpu-type}
11714 A deprecated synonym for @option{-mtune}.
11716 @item -mfpmath=@var{unit}
11718 Generate floating point arithmetics for selected unit @var{unit}. The choices
11719 for @var{unit} are:
11723 Use the standard 387 floating point coprocessor present majority of chips and
11724 emulated otherwise. Code compiled with this option will run almost everywhere.
11725 The temporary results are computed in 80bit precision instead of precision
11726 specified by the type resulting in slightly different results compared to most
11727 of other chips. See @option{-ffloat-store} for more detailed description.
11729 This is the default choice for i386 compiler.
11732 Use scalar floating point instructions present in the SSE instruction set.
11733 This instruction set is supported by Pentium3 and newer chips, in the AMD line
11734 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
11735 instruction set supports only single precision arithmetics, thus the double and
11736 extended precision arithmetics is still done using 387. Later version, present
11737 only in Pentium4 and the future AMD x86-64 chips supports double precision
11740 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
11741 or @option{-msse2} switches to enable SSE extensions and make this option
11742 effective. For the x86-64 compiler, these extensions are enabled by default.
11744 The resulting code should be considerably faster in the majority of cases and avoid
11745 the numerical instability problems of 387 code, but may break some existing
11746 code that expects temporaries to be 80bit.
11748 This is the default choice for the x86-64 compiler.
11753 Attempt to utilize both instruction sets at once. This effectively double the
11754 amount of available registers and on chips with separate execution units for
11755 387 and SSE the execution resources too. Use this option with care, as it is
11756 still experimental, because the GCC register allocator does not model separate
11757 functional units well resulting in instable performance.
11760 @item -masm=@var{dialect}
11761 @opindex masm=@var{dialect}
11762 Output asm instructions using selected @var{dialect}. Supported
11763 choices are @samp{intel} or @samp{att} (the default one). Darwin does
11764 not support @samp{intel}.
11767 @itemx -mno-ieee-fp
11769 @opindex mno-ieee-fp
11770 Control whether or not the compiler uses IEEE floating point
11771 comparisons. These handle correctly the case where the result of a
11772 comparison is unordered.
11775 @opindex msoft-float
11776 Generate output containing library calls for floating point.
11777 @strong{Warning:} the requisite libraries are not part of GCC@.
11778 Normally the facilities of the machine's usual C compiler are used, but
11779 this can't be done directly in cross-compilation. You must make your
11780 own arrangements to provide suitable library functions for
11783 On machines where a function returns floating point results in the 80387
11784 register stack, some floating point opcodes may be emitted even if
11785 @option{-msoft-float} is used.
11787 @item -mno-fp-ret-in-387
11788 @opindex mno-fp-ret-in-387
11789 Do not use the FPU registers for return values of functions.
11791 The usual calling convention has functions return values of types
11792 @code{float} and @code{double} in an FPU register, even if there
11793 is no FPU@. The idea is that the operating system should emulate
11796 The option @option{-mno-fp-ret-in-387} causes such values to be returned
11797 in ordinary CPU registers instead.
11799 @item -mno-fancy-math-387
11800 @opindex mno-fancy-math-387
11801 Some 387 emulators do not support the @code{sin}, @code{cos} and
11802 @code{sqrt} instructions for the 387. Specify this option to avoid
11803 generating those instructions. This option is the default on FreeBSD,
11804 OpenBSD and NetBSD@. This option is overridden when @option{-march}
11805 indicates that the target cpu will always have an FPU and so the
11806 instruction will not need emulation. As of revision 2.6.1, these
11807 instructions are not generated unless you also use the
11808 @option{-funsafe-math-optimizations} switch.
11810 @item -malign-double
11811 @itemx -mno-align-double
11812 @opindex malign-double
11813 @opindex mno-align-double
11814 Control whether GCC aligns @code{double}, @code{long double}, and
11815 @code{long long} variables on a two word boundary or a one word
11816 boundary. Aligning @code{double} variables on a two word boundary will
11817 produce code that runs somewhat faster on a @samp{Pentium} at the
11818 expense of more memory.
11820 On x86-64, @option{-malign-double} is enabled by default.
11822 @strong{Warning:} if you use the @option{-malign-double} switch,
11823 structures containing the above types will be aligned differently than
11824 the published application binary interface specifications for the 386
11825 and will not be binary compatible with structures in code compiled
11826 without that switch.
11828 @item -m96bit-long-double
11829 @itemx -m128bit-long-double
11830 @opindex m96bit-long-double
11831 @opindex m128bit-long-double
11832 These switches control the size of @code{long double} type. The i386
11833 application binary interface specifies the size to be 96 bits,
11834 so @option{-m96bit-long-double} is the default in 32 bit mode.
11836 Modern architectures (Pentium and newer) would prefer @code{long double}
11837 to be aligned to an 8 or 16 byte boundary. In arrays or structures
11838 conforming to the ABI, this would not be possible. So specifying a
11839 @option{-m128bit-long-double} will align @code{long double}
11840 to a 16 byte boundary by padding the @code{long double} with an additional
11843 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
11844 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
11846 Notice that neither of these options enable any extra precision over the x87
11847 standard of 80 bits for a @code{long double}.
11849 @strong{Warning:} if you override the default value for your target ABI, the
11850 structures and arrays containing @code{long double} variables will change
11851 their size as well as function calling convention for function taking
11852 @code{long double} will be modified. Hence they will not be binary
11853 compatible with arrays or structures in code compiled without that switch.
11855 @item -mlarge-data-threshold=@var{number}
11856 @opindex mlarge-data-threshold=@var{number}
11857 When @option{-mcmodel=medium} is specified, the data greater than
11858 @var{threshold} are placed in large data section. This value must be the
11859 same across all object linked into the binary and defaults to 65535.
11863 Use a different function-calling convention, in which functions that
11864 take a fixed number of arguments return with the @code{ret} @var{num}
11865 instruction, which pops their arguments while returning. This saves one
11866 instruction in the caller since there is no need to pop the arguments
11869 You can specify that an individual function is called with this calling
11870 sequence with the function attribute @samp{stdcall}. You can also
11871 override the @option{-mrtd} option by using the function attribute
11872 @samp{cdecl}. @xref{Function Attributes}.
11874 @strong{Warning:} this calling convention is incompatible with the one
11875 normally used on Unix, so you cannot use it if you need to call
11876 libraries compiled with the Unix compiler.
11878 Also, you must provide function prototypes for all functions that
11879 take variable numbers of arguments (including @code{printf});
11880 otherwise incorrect code will be generated for calls to those
11883 In addition, seriously incorrect code will result if you call a
11884 function with too many arguments. (Normally, extra arguments are
11885 harmlessly ignored.)
11887 @item -mregparm=@var{num}
11889 Control how many registers are used to pass integer arguments. By
11890 default, no registers are used to pass arguments, and at most 3
11891 registers can be used. You can control this behavior for a specific
11892 function by using the function attribute @samp{regparm}.
11893 @xref{Function Attributes}.
11895 @strong{Warning:} if you use this switch, and
11896 @var{num} is nonzero, then you must build all modules with the same
11897 value, including any libraries. This includes the system libraries and
11901 @opindex msseregparm
11902 Use SSE register passing conventions for float and double arguments
11903 and return values. You can control this behavior for a specific
11904 function by using the function attribute @samp{sseregparm}.
11905 @xref{Function Attributes}.
11907 @strong{Warning:} if you use this switch then you must build all
11908 modules with the same value, including any libraries. This includes
11909 the system libraries and startup modules.
11918 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
11919 is specified, the significands of results of floating-point operations are
11920 rounded to 24 bits (single precision); @option{-mpc64} rounds the
11921 significands of results of floating-point operations to 53 bits (double
11922 precision) and @option{-mpc80} rounds the significands of results of
11923 floating-point operations to 64 bits (extended double precision), which is
11924 the default. When this option is used, floating-point operations in higher
11925 precisions are not available to the programmer without setting the FPU
11926 control word explicitly.
11928 Setting the rounding of floating-point operations to less than the default
11929 80 bits can speed some programs by 2% or more. Note that some mathematical
11930 libraries assume that extended precision (80 bit) floating-point operations
11931 are enabled by default; routines in such libraries could suffer significant
11932 loss of accuracy, typically through so-called "catastrophic cancellation",
11933 when this option is used to set the precision to less than extended precision.
11935 @item -mstackrealign
11936 @opindex mstackrealign
11937 Realign the stack at entry. On the Intel x86, the @option{-mstackrealign}
11938 option will generate an alternate prologue and epilogue that realigns the
11939 runtime stack if necessary. This supports mixing legacy codes that keep
11940 a 4-byte aligned stack with modern codes that keep a 16-byte stack for
11941 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
11942 applicable to individual functions.
11944 @item -mpreferred-stack-boundary=@var{num}
11945 @opindex mpreferred-stack-boundary
11946 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
11947 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
11948 the default is 4 (16 bytes or 128 bits).
11950 @item -mincoming-stack-boundary=@var{num}
11951 @opindex mincoming-stack-boundary
11952 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
11953 boundary. If @option{-mincoming-stack-boundary} is not specified,
11954 the one specified by @option{-mpreferred-stack-boundary} will be used.
11956 On Pentium and PentiumPro, @code{double} and @code{long double} values
11957 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
11958 suffer significant run time performance penalties. On Pentium III, the
11959 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
11960 properly if it is not 16 byte aligned.
11962 To ensure proper alignment of this values on the stack, the stack boundary
11963 must be as aligned as that required by any value stored on the stack.
11964 Further, every function must be generated such that it keeps the stack
11965 aligned. Thus calling a function compiled with a higher preferred
11966 stack boundary from a function compiled with a lower preferred stack
11967 boundary will most likely misalign the stack. It is recommended that
11968 libraries that use callbacks always use the default setting.
11970 This extra alignment does consume extra stack space, and generally
11971 increases code size. Code that is sensitive to stack space usage, such
11972 as embedded systems and operating system kernels, may want to reduce the
11973 preferred alignment to @option{-mpreferred-stack-boundary=2}.
12013 These switches enable or disable the use of instructions in the MMX,
12014 SSE, SSE2, SSE3, SSSE3, SSE4.1, AVX, AES, PCLMUL, SSE4A, FMA4, ABM or
12015 3DNow!@: extended instruction sets.
12016 These extensions are also available as built-in functions: see
12017 @ref{X86 Built-in Functions}, for details of the functions enabled and
12018 disabled by these switches.
12020 To have SSE/SSE2 instructions generated automatically from floating-point
12021 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
12023 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
12024 generates new AVX instructions or AVX equivalence for all SSEx instructions
12027 These options will enable GCC to use these extended instructions in
12028 generated code, even without @option{-mfpmath=sse}. Applications which
12029 perform runtime CPU detection must compile separate files for each
12030 supported architecture, using the appropriate flags. In particular,
12031 the file containing the CPU detection code should be compiled without
12036 This option instructs GCC to emit a @code{cld} instruction in the prologue
12037 of functions that use string instructions. String instructions depend on
12038 the DF flag to select between autoincrement or autodecrement mode. While the
12039 ABI specifies the DF flag to be cleared on function entry, some operating
12040 systems violate this specification by not clearing the DF flag in their
12041 exception dispatchers. The exception handler can be invoked with the DF flag
12042 set which leads to wrong direction mode, when string instructions are used.
12043 This option can be enabled by default on 32-bit x86 targets by configuring
12044 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
12045 instructions can be suppressed with the @option{-mno-cld} compiler option
12050 This option will enable GCC to use CMPXCHG16B instruction in generated code.
12051 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
12052 data types. This is useful for high resolution counters that could be updated
12053 by multiple processors (or cores). This instruction is generated as part of
12054 atomic built-in functions: see @ref{Atomic Builtins} for details.
12058 This option will enable GCC to use SAHF instruction in generated 64-bit code.
12059 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
12060 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
12061 SAHF are load and store instructions, respectively, for certain status flags.
12062 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
12063 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
12067 This option will enable GCC to use movbe instruction to implement
12068 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
12072 This option will enable built-in functions, @code{__builtin_ia32_crc32qi},
12073 @code{__builtin_ia32_crc32hi}. @code{__builtin_ia32_crc32si} and
12074 @code{__builtin_ia32_crc32di} to generate the crc32 machine instruction.
12078 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
12079 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Raphson step
12080 to increase precision instead of DIVSS and SQRTSS (and their vectorized
12081 variants) for single precision floating point arguments. These instructions
12082 are generated only when @option{-funsafe-math-optimizations} is enabled
12083 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
12084 Note that while the throughput of the sequence is higher than the throughput
12085 of the non-reciprocal instruction, the precision of the sequence can be
12086 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
12088 @item -mveclibabi=@var{type}
12089 @opindex mveclibabi
12090 Specifies the ABI type to use for vectorizing intrinsics using an
12091 external library. Supported types are @code{svml} for the Intel short
12092 vector math library and @code{acml} for the AMD math core library style
12093 of interfacing. GCC will currently emit calls to @code{vmldExp2},
12094 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
12095 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
12096 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
12097 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
12098 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
12099 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
12100 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
12101 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
12102 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
12103 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
12104 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
12105 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
12106 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
12107 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
12108 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
12109 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
12110 compatible library will have to be specified at link time.
12112 @item -mabi=@var{name}
12114 Generate code for the specified calling convention. Permissible values
12115 are: @samp{sysv} for the ABI used on GNU/Linux and other systems and
12116 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
12117 ABI when targeting Windows. On all other systems, the default is the
12118 SYSV ABI. You can control this behavior for a specific function by
12119 using the function attribute @samp{ms_abi}/@samp{sysv_abi}.
12120 @xref{Function Attributes}.
12123 @itemx -mno-push-args
12124 @opindex mpush-args
12125 @opindex mno-push-args
12126 Use PUSH operations to store outgoing parameters. This method is shorter
12127 and usually equally fast as method using SUB/MOV operations and is enabled
12128 by default. In some cases disabling it may improve performance because of
12129 improved scheduling and reduced dependencies.
12131 @item -maccumulate-outgoing-args
12132 @opindex maccumulate-outgoing-args
12133 If enabled, the maximum amount of space required for outgoing arguments will be
12134 computed in the function prologue. This is faster on most modern CPUs
12135 because of reduced dependencies, improved scheduling and reduced stack usage
12136 when preferred stack boundary is not equal to 2. The drawback is a notable
12137 increase in code size. This switch implies @option{-mno-push-args}.
12141 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
12142 on thread-safe exception handling must compile and link all code with the
12143 @option{-mthreads} option. When compiling, @option{-mthreads} defines
12144 @option{-D_MT}; when linking, it links in a special thread helper library
12145 @option{-lmingwthrd} which cleans up per thread exception handling data.
12147 @item -mno-align-stringops
12148 @opindex mno-align-stringops
12149 Do not align destination of inlined string operations. This switch reduces
12150 code size and improves performance in case the destination is already aligned,
12151 but GCC doesn't know about it.
12153 @item -minline-all-stringops
12154 @opindex minline-all-stringops
12155 By default GCC inlines string operations only when destination is known to be
12156 aligned at least to 4 byte boundary. This enables more inlining, increase code
12157 size, but may improve performance of code that depends on fast memcpy, strlen
12158 and memset for short lengths.
12160 @item -minline-stringops-dynamically
12161 @opindex minline-stringops-dynamically
12162 For string operation of unknown size, inline runtime checks so for small
12163 blocks inline code is used, while for large blocks library call is used.
12165 @item -mstringop-strategy=@var{alg}
12166 @opindex mstringop-strategy=@var{alg}
12167 Overwrite internal decision heuristic about particular algorithm to inline
12168 string operation with. The allowed values are @code{rep_byte},
12169 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
12170 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
12171 expanding inline loop, @code{libcall} for always expanding library call.
12173 @item -momit-leaf-frame-pointer
12174 @opindex momit-leaf-frame-pointer
12175 Don't keep the frame pointer in a register for leaf functions. This
12176 avoids the instructions to save, set up and restore frame pointers and
12177 makes an extra register available in leaf functions. The option
12178 @option{-fomit-frame-pointer} removes the frame pointer for all functions
12179 which might make debugging harder.
12181 @item -mtls-direct-seg-refs
12182 @itemx -mno-tls-direct-seg-refs
12183 @opindex mtls-direct-seg-refs
12184 Controls whether TLS variables may be accessed with offsets from the
12185 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
12186 or whether the thread base pointer must be added. Whether or not this
12187 is legal depends on the operating system, and whether it maps the
12188 segment to cover the entire TLS area.
12190 For systems that use GNU libc, the default is on.
12193 @itemx -mno-sse2avx
12195 Specify that the assembler should encode SSE instructions with VEX
12196 prefix. The option @option{-mavx} turns this on by default.
12199 These @samp{-m} switches are supported in addition to the above
12200 on AMD x86-64 processors in 64-bit environments.
12207 Generate code for a 32-bit or 64-bit environment.
12208 The 32-bit environment sets int, long and pointer to 32 bits and
12209 generates code that runs on any i386 system.
12210 The 64-bit environment sets int to 32 bits and long and pointer
12211 to 64 bits and generates code for AMD's x86-64 architecture. For
12212 darwin only the -m64 option turns off the @option{-fno-pic} and
12213 @option{-mdynamic-no-pic} options.
12215 @item -mno-red-zone
12216 @opindex mno-red-zone
12217 Do not use a so called red zone for x86-64 code. The red zone is mandated
12218 by the x86-64 ABI, it is a 128-byte area beyond the location of the
12219 stack pointer that will not be modified by signal or interrupt handlers
12220 and therefore can be used for temporary data without adjusting the stack
12221 pointer. The flag @option{-mno-red-zone} disables this red zone.
12223 @item -mcmodel=small
12224 @opindex mcmodel=small
12225 Generate code for the small code model: the program and its symbols must
12226 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
12227 Programs can be statically or dynamically linked. This is the default
12230 @item -mcmodel=kernel
12231 @opindex mcmodel=kernel
12232 Generate code for the kernel code model. The kernel runs in the
12233 negative 2 GB of the address space.
12234 This model has to be used for Linux kernel code.
12236 @item -mcmodel=medium
12237 @opindex mcmodel=medium
12238 Generate code for the medium model: The program is linked in the lower 2
12239 GB of the address space. Small symbols are also placed there. Symbols
12240 with sizes larger than @option{-mlarge-data-threshold} are put into
12241 large data or bss sections and can be located above 2GB. Programs can
12242 be statically or dynamically linked.
12244 @item -mcmodel=large
12245 @opindex mcmodel=large
12246 Generate code for the large model: This model makes no assumptions
12247 about addresses and sizes of sections.
12250 @node IA-64 Options
12251 @subsection IA-64 Options
12252 @cindex IA-64 Options
12254 These are the @samp{-m} options defined for the Intel IA-64 architecture.
12258 @opindex mbig-endian
12259 Generate code for a big endian target. This is the default for HP-UX@.
12261 @item -mlittle-endian
12262 @opindex mlittle-endian
12263 Generate code for a little endian target. This is the default for AIX5
12269 @opindex mno-gnu-as
12270 Generate (or don't) code for the GNU assembler. This is the default.
12271 @c Also, this is the default if the configure option @option{--with-gnu-as}
12277 @opindex mno-gnu-ld
12278 Generate (or don't) code for the GNU linker. This is the default.
12279 @c Also, this is the default if the configure option @option{--with-gnu-ld}
12284 Generate code that does not use a global pointer register. The result
12285 is not position independent code, and violates the IA-64 ABI@.
12287 @item -mvolatile-asm-stop
12288 @itemx -mno-volatile-asm-stop
12289 @opindex mvolatile-asm-stop
12290 @opindex mno-volatile-asm-stop
12291 Generate (or don't) a stop bit immediately before and after volatile asm
12294 @item -mregister-names
12295 @itemx -mno-register-names
12296 @opindex mregister-names
12297 @opindex mno-register-names
12298 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
12299 the stacked registers. This may make assembler output more readable.
12305 Disable (or enable) optimizations that use the small data section. This may
12306 be useful for working around optimizer bugs.
12308 @item -mconstant-gp
12309 @opindex mconstant-gp
12310 Generate code that uses a single constant global pointer value. This is
12311 useful when compiling kernel code.
12315 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
12316 This is useful when compiling firmware code.
12318 @item -minline-float-divide-min-latency
12319 @opindex minline-float-divide-min-latency
12320 Generate code for inline divides of floating point values
12321 using the minimum latency algorithm.
12323 @item -minline-float-divide-max-throughput
12324 @opindex minline-float-divide-max-throughput
12325 Generate code for inline divides of floating point values
12326 using the maximum throughput algorithm.
12328 @item -mno-inline-float-divide
12329 @opindex mno-inline-float-divide
12330 Do not generate inline code for divides of floating point values.
12332 @item -minline-int-divide-min-latency
12333 @opindex minline-int-divide-min-latency
12334 Generate code for inline divides of integer values
12335 using the minimum latency algorithm.
12337 @item -minline-int-divide-max-throughput
12338 @opindex minline-int-divide-max-throughput
12339 Generate code for inline divides of integer values
12340 using the maximum throughput algorithm.
12342 @item -mno-inline-int-divide
12343 @opindex mno-inline-int-divide
12344 Do not generate inline code for divides of integer values.
12346 @item -minline-sqrt-min-latency
12347 @opindex minline-sqrt-min-latency
12348 Generate code for inline square roots
12349 using the minimum latency algorithm.
12351 @item -minline-sqrt-max-throughput
12352 @opindex minline-sqrt-max-throughput
12353 Generate code for inline square roots
12354 using the maximum throughput algorithm.
12356 @item -mno-inline-sqrt
12357 @opindex mno-inline-sqrt
12358 Do not generate inline code for sqrt.
12361 @itemx -mno-fused-madd
12362 @opindex mfused-madd
12363 @opindex mno-fused-madd
12364 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12365 instructions. The default is to use these instructions.
12367 @item -mno-dwarf2-asm
12368 @itemx -mdwarf2-asm
12369 @opindex mno-dwarf2-asm
12370 @opindex mdwarf2-asm
12371 Don't (or do) generate assembler code for the DWARF2 line number debugging
12372 info. This may be useful when not using the GNU assembler.
12374 @item -mearly-stop-bits
12375 @itemx -mno-early-stop-bits
12376 @opindex mearly-stop-bits
12377 @opindex mno-early-stop-bits
12378 Allow stop bits to be placed earlier than immediately preceding the
12379 instruction that triggered the stop bit. This can improve instruction
12380 scheduling, but does not always do so.
12382 @item -mfixed-range=@var{register-range}
12383 @opindex mfixed-range
12384 Generate code treating the given register range as fixed registers.
12385 A fixed register is one that the register allocator can not use. This is
12386 useful when compiling kernel code. A register range is specified as
12387 two registers separated by a dash. Multiple register ranges can be
12388 specified separated by a comma.
12390 @item -mtls-size=@var{tls-size}
12392 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
12395 @item -mtune=@var{cpu-type}
12397 Tune the instruction scheduling for a particular CPU, Valid values are
12398 itanium, itanium1, merced, itanium2, and mckinley.
12404 Generate code for a 32-bit or 64-bit environment.
12405 The 32-bit environment sets int, long and pointer to 32 bits.
12406 The 64-bit environment sets int to 32 bits and long and pointer
12407 to 64 bits. These are HP-UX specific flags.
12409 @item -mno-sched-br-data-spec
12410 @itemx -msched-br-data-spec
12411 @opindex mno-sched-br-data-spec
12412 @opindex msched-br-data-spec
12413 (Dis/En)able data speculative scheduling before reload.
12414 This will result in generation of the ld.a instructions and
12415 the corresponding check instructions (ld.c / chk.a).
12416 The default is 'disable'.
12418 @item -msched-ar-data-spec
12419 @itemx -mno-sched-ar-data-spec
12420 @opindex msched-ar-data-spec
12421 @opindex mno-sched-ar-data-spec
12422 (En/Dis)able data speculative scheduling after reload.
12423 This will result in generation of the ld.a instructions and
12424 the corresponding check instructions (ld.c / chk.a).
12425 The default is 'enable'.
12427 @item -mno-sched-control-spec
12428 @itemx -msched-control-spec
12429 @opindex mno-sched-control-spec
12430 @opindex msched-control-spec
12431 (Dis/En)able control speculative scheduling. This feature is
12432 available only during region scheduling (i.e.@: before reload).
12433 This will result in generation of the ld.s instructions and
12434 the corresponding check instructions chk.s .
12435 The default is 'disable'.
12437 @item -msched-br-in-data-spec
12438 @itemx -mno-sched-br-in-data-spec
12439 @opindex msched-br-in-data-spec
12440 @opindex mno-sched-br-in-data-spec
12441 (En/Dis)able speculative scheduling of the instructions that
12442 are dependent on the data speculative loads before reload.
12443 This is effective only with @option{-msched-br-data-spec} enabled.
12444 The default is 'enable'.
12446 @item -msched-ar-in-data-spec
12447 @itemx -mno-sched-ar-in-data-spec
12448 @opindex msched-ar-in-data-spec
12449 @opindex mno-sched-ar-in-data-spec
12450 (En/Dis)able speculative scheduling of the instructions that
12451 are dependent on the data speculative loads after reload.
12452 This is effective only with @option{-msched-ar-data-spec} enabled.
12453 The default is 'enable'.
12455 @item -msched-in-control-spec
12456 @itemx -mno-sched-in-control-spec
12457 @opindex msched-in-control-spec
12458 @opindex mno-sched-in-control-spec
12459 (En/Dis)able speculative scheduling of the instructions that
12460 are dependent on the control speculative loads.
12461 This is effective only with @option{-msched-control-spec} enabled.
12462 The default is 'enable'.
12464 @item -mno-sched-prefer-non-data-spec-insns
12465 @itemx -msched-prefer-non-data-spec-insns
12466 @opindex mno-sched-prefer-non-data-spec-insns
12467 @opindex msched-prefer-non-data-spec-insns
12468 If enabled, data speculative instructions will be chosen for schedule
12469 only if there are no other choices at the moment. This will make
12470 the use of the data speculation much more conservative.
12471 The default is 'disable'.
12473 @item -mno-sched-prefer-non-control-spec-insns
12474 @itemx -msched-prefer-non-control-spec-insns
12475 @opindex mno-sched-prefer-non-control-spec-insns
12476 @opindex msched-prefer-non-control-spec-insns
12477 If enabled, control speculative instructions will be chosen for schedule
12478 only if there are no other choices at the moment. This will make
12479 the use of the control speculation much more conservative.
12480 The default is 'disable'.
12482 @item -mno-sched-count-spec-in-critical-path
12483 @itemx -msched-count-spec-in-critical-path
12484 @opindex mno-sched-count-spec-in-critical-path
12485 @opindex msched-count-spec-in-critical-path
12486 If enabled, speculative dependencies will be considered during
12487 computation of the instructions priorities. This will make the use of the
12488 speculation a bit more conservative.
12489 The default is 'disable'.
12491 @item -msched-spec-ldc
12492 @opindex msched-spec-ldc
12493 Use a simple data speculation check. This option is on by default.
12495 @item -msched-control-spec-ldc
12496 @opindex msched-spec-ldc
12497 Use a simple check for control speculation. This option is on by default.
12499 @item -msched-stop-bits-after-every-cycle
12500 @opindex msched-stop-bits-after-every-cycle
12501 Place a stop bit after every cycle when scheduling. This option is on
12504 @item -msched-fp-mem-deps-zero-cost
12505 @opindex msched-fp-mem-deps-zero-cost
12506 Assume that floating-point stores and loads are not likely to cause a conflict
12507 when placed into the same instruction group. This option is disabled by
12510 @item -msel-sched-dont-check-control-spec
12511 @opindex msel-sched-dont-check-control-spec
12512 Generate checks for control speculation in selective scheduling.
12513 This flag is disabled by default.
12515 @item -msched-max-memory-insns=@var{max-insns}
12516 @opindex msched-max-memory-insns
12517 Limit on the number of memory insns per instruction group, giving lower
12518 priority to subsequent memory insns attempting to schedule in the same
12519 instruction group. Frequently useful to prevent cache bank conflicts.
12520 The default value is 1.
12522 @item -msched-max-memory-insns-hard-limit
12523 @opindex msched-max-memory-insns-hard-limit
12524 Disallow more than `msched-max-memory-insns' in instruction group.
12525 Otherwise, limit is `soft' meaning that we would prefer non-memory operations
12526 when limit is reached but may still schedule memory operations.
12530 @node IA-64/VMS Options
12531 @subsection IA-64/VMS Options
12533 These @samp{-m} options are defined for the IA-64/VMS implementations:
12536 @item -mvms-return-codes
12537 @opindex mvms-return-codes
12538 Return VMS condition codes from main. The default is to return POSIX
12539 style condition (e.g.@ error) codes.
12541 @item -mdebug-main=@var{prefix}
12542 @opindex mdebug-main=@var{prefix}
12543 Flag the first routine whose name starts with @var{prefix} as the main
12544 routine for the debugger.
12548 Default to 64bit memory allocation routines.
12552 @subsection M32C Options
12553 @cindex M32C options
12556 @item -mcpu=@var{name}
12558 Select the CPU for which code is generated. @var{name} may be one of
12559 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
12560 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
12561 the M32C/80 series.
12565 Specifies that the program will be run on the simulator. This causes
12566 an alternate runtime library to be linked in which supports, for
12567 example, file I/O@. You must not use this option when generating
12568 programs that will run on real hardware; you must provide your own
12569 runtime library for whatever I/O functions are needed.
12571 @item -memregs=@var{number}
12573 Specifies the number of memory-based pseudo-registers GCC will use
12574 during code generation. These pseudo-registers will be used like real
12575 registers, so there is a tradeoff between GCC's ability to fit the
12576 code into available registers, and the performance penalty of using
12577 memory instead of registers. Note that all modules in a program must
12578 be compiled with the same value for this option. Because of that, you
12579 must not use this option with the default runtime libraries gcc
12584 @node M32R/D Options
12585 @subsection M32R/D Options
12586 @cindex M32R/D options
12588 These @option{-m} options are defined for Renesas M32R/D architectures:
12593 Generate code for the M32R/2@.
12597 Generate code for the M32R/X@.
12601 Generate code for the M32R@. This is the default.
12603 @item -mmodel=small
12604 @opindex mmodel=small
12605 Assume all objects live in the lower 16MB of memory (so that their addresses
12606 can be loaded with the @code{ld24} instruction), and assume all subroutines
12607 are reachable with the @code{bl} instruction.
12608 This is the default.
12610 The addressability of a particular object can be set with the
12611 @code{model} attribute.
12613 @item -mmodel=medium
12614 @opindex mmodel=medium
12615 Assume objects may be anywhere in the 32-bit address space (the compiler
12616 will generate @code{seth/add3} instructions to load their addresses), and
12617 assume all subroutines are reachable with the @code{bl} instruction.
12619 @item -mmodel=large
12620 @opindex mmodel=large
12621 Assume objects may be anywhere in the 32-bit address space (the compiler
12622 will generate @code{seth/add3} instructions to load their addresses), and
12623 assume subroutines may not be reachable with the @code{bl} instruction
12624 (the compiler will generate the much slower @code{seth/add3/jl}
12625 instruction sequence).
12628 @opindex msdata=none
12629 Disable use of the small data area. Variables will be put into
12630 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
12631 @code{section} attribute has been specified).
12632 This is the default.
12634 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
12635 Objects may be explicitly put in the small data area with the
12636 @code{section} attribute using one of these sections.
12638 @item -msdata=sdata
12639 @opindex msdata=sdata
12640 Put small global and static data in the small data area, but do not
12641 generate special code to reference them.
12644 @opindex msdata=use
12645 Put small global and static data in the small data area, and generate
12646 special instructions to reference them.
12650 @cindex smaller data references
12651 Put global and static objects less than or equal to @var{num} bytes
12652 into the small data or bss sections instead of the normal data or bss
12653 sections. The default value of @var{num} is 8.
12654 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
12655 for this option to have any effect.
12657 All modules should be compiled with the same @option{-G @var{num}} value.
12658 Compiling with different values of @var{num} may or may not work; if it
12659 doesn't the linker will give an error message---incorrect code will not be
12664 Makes the M32R specific code in the compiler display some statistics
12665 that might help in debugging programs.
12667 @item -malign-loops
12668 @opindex malign-loops
12669 Align all loops to a 32-byte boundary.
12671 @item -mno-align-loops
12672 @opindex mno-align-loops
12673 Do not enforce a 32-byte alignment for loops. This is the default.
12675 @item -missue-rate=@var{number}
12676 @opindex missue-rate=@var{number}
12677 Issue @var{number} instructions per cycle. @var{number} can only be 1
12680 @item -mbranch-cost=@var{number}
12681 @opindex mbranch-cost=@var{number}
12682 @var{number} can only be 1 or 2. If it is 1 then branches will be
12683 preferred over conditional code, if it is 2, then the opposite will
12686 @item -mflush-trap=@var{number}
12687 @opindex mflush-trap=@var{number}
12688 Specifies the trap number to use to flush the cache. The default is
12689 12. Valid numbers are between 0 and 15 inclusive.
12691 @item -mno-flush-trap
12692 @opindex mno-flush-trap
12693 Specifies that the cache cannot be flushed by using a trap.
12695 @item -mflush-func=@var{name}
12696 @opindex mflush-func=@var{name}
12697 Specifies the name of the operating system function to call to flush
12698 the cache. The default is @emph{_flush_cache}, but a function call
12699 will only be used if a trap is not available.
12701 @item -mno-flush-func
12702 @opindex mno-flush-func
12703 Indicates that there is no OS function for flushing the cache.
12707 @node M680x0 Options
12708 @subsection M680x0 Options
12709 @cindex M680x0 options
12711 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
12712 The default settings depend on which architecture was selected when
12713 the compiler was configured; the defaults for the most common choices
12717 @item -march=@var{arch}
12719 Generate code for a specific M680x0 or ColdFire instruction set
12720 architecture. Permissible values of @var{arch} for M680x0
12721 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
12722 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
12723 architectures are selected according to Freescale's ISA classification
12724 and the permissible values are: @samp{isaa}, @samp{isaaplus},
12725 @samp{isab} and @samp{isac}.
12727 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
12728 code for a ColdFire target. The @var{arch} in this macro is one of the
12729 @option{-march} arguments given above.
12731 When used together, @option{-march} and @option{-mtune} select code
12732 that runs on a family of similar processors but that is optimized
12733 for a particular microarchitecture.
12735 @item -mcpu=@var{cpu}
12737 Generate code for a specific M680x0 or ColdFire processor.
12738 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
12739 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
12740 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
12741 below, which also classifies the CPUs into families:
12743 @multitable @columnfractions 0.20 0.80
12744 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
12745 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51cn} @samp{51em} @samp{51qe}
12746 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
12747 @item @samp{5206e} @tab @samp{5206e}
12748 @item @samp{5208} @tab @samp{5207} @samp{5208}
12749 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
12750 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
12751 @item @samp{5216} @tab @samp{5214} @samp{5216}
12752 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
12753 @item @samp{5225} @tab @samp{5224} @samp{5225}
12754 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
12755 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
12756 @item @samp{5249} @tab @samp{5249}
12757 @item @samp{5250} @tab @samp{5250}
12758 @item @samp{5271} @tab @samp{5270} @samp{5271}
12759 @item @samp{5272} @tab @samp{5272}
12760 @item @samp{5275} @tab @samp{5274} @samp{5275}
12761 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
12762 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
12763 @item @samp{5307} @tab @samp{5307}
12764 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
12765 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
12766 @item @samp{5407} @tab @samp{5407}
12767 @item @samp{5475} @tab @samp{5470} @samp{5471} @samp{5472} @samp{5473} @samp{5474} @samp{5475} @samp{547x} @samp{5480} @samp{5481} @samp{5482} @samp{5483} @samp{5484} @samp{5485}
12770 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
12771 @var{arch} is compatible with @var{cpu}. Other combinations of
12772 @option{-mcpu} and @option{-march} are rejected.
12774 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
12775 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
12776 where the value of @var{family} is given by the table above.
12778 @item -mtune=@var{tune}
12780 Tune the code for a particular microarchitecture, within the
12781 constraints set by @option{-march} and @option{-mcpu}.
12782 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
12783 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
12784 and @samp{cpu32}. The ColdFire microarchitectures
12785 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
12787 You can also use @option{-mtune=68020-40} for code that needs
12788 to run relatively well on 68020, 68030 and 68040 targets.
12789 @option{-mtune=68020-60} is similar but includes 68060 targets
12790 as well. These two options select the same tuning decisions as
12791 @option{-m68020-40} and @option{-m68020-60} respectively.
12793 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
12794 when tuning for 680x0 architecture @var{arch}. It also defines
12795 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
12796 option is used. If gcc is tuning for a range of architectures,
12797 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
12798 it defines the macros for every architecture in the range.
12800 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
12801 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
12802 of the arguments given above.
12808 Generate output for a 68000. This is the default
12809 when the compiler is configured for 68000-based systems.
12810 It is equivalent to @option{-march=68000}.
12812 Use this option for microcontrollers with a 68000 or EC000 core,
12813 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
12817 Generate output for a 68010. This is the default
12818 when the compiler is configured for 68010-based systems.
12819 It is equivalent to @option{-march=68010}.
12825 Generate output for a 68020. This is the default
12826 when the compiler is configured for 68020-based systems.
12827 It is equivalent to @option{-march=68020}.
12831 Generate output for a 68030. This is the default when the compiler is
12832 configured for 68030-based systems. It is equivalent to
12833 @option{-march=68030}.
12837 Generate output for a 68040. This is the default when the compiler is
12838 configured for 68040-based systems. It is equivalent to
12839 @option{-march=68040}.
12841 This option inhibits the use of 68881/68882 instructions that have to be
12842 emulated by software on the 68040. Use this option if your 68040 does not
12843 have code to emulate those instructions.
12847 Generate output for a 68060. This is the default when the compiler is
12848 configured for 68060-based systems. It is equivalent to
12849 @option{-march=68060}.
12851 This option inhibits the use of 68020 and 68881/68882 instructions that
12852 have to be emulated by software on the 68060. Use this option if your 68060
12853 does not have code to emulate those instructions.
12857 Generate output for a CPU32. This is the default
12858 when the compiler is configured for CPU32-based systems.
12859 It is equivalent to @option{-march=cpu32}.
12861 Use this option for microcontrollers with a
12862 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
12863 68336, 68340, 68341, 68349 and 68360.
12867 Generate output for a 520X ColdFire CPU@. This is the default
12868 when the compiler is configured for 520X-based systems.
12869 It is equivalent to @option{-mcpu=5206}, and is now deprecated
12870 in favor of that option.
12872 Use this option for microcontroller with a 5200 core, including
12873 the MCF5202, MCF5203, MCF5204 and MCF5206.
12877 Generate output for a 5206e ColdFire CPU@. The option is now
12878 deprecated in favor of the equivalent @option{-mcpu=5206e}.
12882 Generate output for a member of the ColdFire 528X family.
12883 The option is now deprecated in favor of the equivalent
12884 @option{-mcpu=528x}.
12888 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
12889 in favor of the equivalent @option{-mcpu=5307}.
12893 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
12894 in favor of the equivalent @option{-mcpu=5407}.
12898 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
12899 This includes use of hardware floating point instructions.
12900 The option is equivalent to @option{-mcpu=547x}, and is now
12901 deprecated in favor of that option.
12905 Generate output for a 68040, without using any of the new instructions.
12906 This results in code which can run relatively efficiently on either a
12907 68020/68881 or a 68030 or a 68040. The generated code does use the
12908 68881 instructions that are emulated on the 68040.
12910 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
12914 Generate output for a 68060, without using any of the new instructions.
12915 This results in code which can run relatively efficiently on either a
12916 68020/68881 or a 68030 or a 68040. The generated code does use the
12917 68881 instructions that are emulated on the 68060.
12919 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
12923 @opindex mhard-float
12925 Generate floating-point instructions. This is the default for 68020
12926 and above, and for ColdFire devices that have an FPU@. It defines the
12927 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
12928 on ColdFire targets.
12931 @opindex msoft-float
12932 Do not generate floating-point instructions; use library calls instead.
12933 This is the default for 68000, 68010, and 68832 targets. It is also
12934 the default for ColdFire devices that have no FPU.
12940 Generate (do not generate) ColdFire hardware divide and remainder
12941 instructions. If @option{-march} is used without @option{-mcpu},
12942 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
12943 architectures. Otherwise, the default is taken from the target CPU
12944 (either the default CPU, or the one specified by @option{-mcpu}). For
12945 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
12946 @option{-mcpu=5206e}.
12948 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
12952 Consider type @code{int} to be 16 bits wide, like @code{short int}.
12953 Additionally, parameters passed on the stack are also aligned to a
12954 16-bit boundary even on targets whose API mandates promotion to 32-bit.
12958 Do not consider type @code{int} to be 16 bits wide. This is the default.
12961 @itemx -mno-bitfield
12962 @opindex mnobitfield
12963 @opindex mno-bitfield
12964 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
12965 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
12969 Do use the bit-field instructions. The @option{-m68020} option implies
12970 @option{-mbitfield}. This is the default if you use a configuration
12971 designed for a 68020.
12975 Use a different function-calling convention, in which functions
12976 that take a fixed number of arguments return with the @code{rtd}
12977 instruction, which pops their arguments while returning. This
12978 saves one instruction in the caller since there is no need to pop
12979 the arguments there.
12981 This calling convention is incompatible with the one normally
12982 used on Unix, so you cannot use it if you need to call libraries
12983 compiled with the Unix compiler.
12985 Also, you must provide function prototypes for all functions that
12986 take variable numbers of arguments (including @code{printf});
12987 otherwise incorrect code will be generated for calls to those
12990 In addition, seriously incorrect code will result if you call a
12991 function with too many arguments. (Normally, extra arguments are
12992 harmlessly ignored.)
12994 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
12995 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
12999 Do not use the calling conventions selected by @option{-mrtd}.
13000 This is the default.
13003 @itemx -mno-align-int
13004 @opindex malign-int
13005 @opindex mno-align-int
13006 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
13007 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
13008 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
13009 Aligning variables on 32-bit boundaries produces code that runs somewhat
13010 faster on processors with 32-bit busses at the expense of more memory.
13012 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
13013 align structures containing the above types differently than
13014 most published application binary interface specifications for the m68k.
13018 Use the pc-relative addressing mode of the 68000 directly, instead of
13019 using a global offset table. At present, this option implies @option{-fpic},
13020 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
13021 not presently supported with @option{-mpcrel}, though this could be supported for
13022 68020 and higher processors.
13024 @item -mno-strict-align
13025 @itemx -mstrict-align
13026 @opindex mno-strict-align
13027 @opindex mstrict-align
13028 Do not (do) assume that unaligned memory references will be handled by
13032 Generate code that allows the data segment to be located in a different
13033 area of memory from the text segment. This allows for execute in place in
13034 an environment without virtual memory management. This option implies
13037 @item -mno-sep-data
13038 Generate code that assumes that the data segment follows the text segment.
13039 This is the default.
13041 @item -mid-shared-library
13042 Generate code that supports shared libraries via the library ID method.
13043 This allows for execute in place and shared libraries in an environment
13044 without virtual memory management. This option implies @option{-fPIC}.
13046 @item -mno-id-shared-library
13047 Generate code that doesn't assume ID based shared libraries are being used.
13048 This is the default.
13050 @item -mshared-library-id=n
13051 Specified the identification number of the ID based shared library being
13052 compiled. Specifying a value of 0 will generate more compact code, specifying
13053 other values will force the allocation of that number to the current
13054 library but is no more space or time efficient than omitting this option.
13060 When generating position-independent code for ColdFire, generate code
13061 that works if the GOT has more than 8192 entries. This code is
13062 larger and slower than code generated without this option. On M680x0
13063 processors, this option is not needed; @option{-fPIC} suffices.
13065 GCC normally uses a single instruction to load values from the GOT@.
13066 While this is relatively efficient, it only works if the GOT
13067 is smaller than about 64k. Anything larger causes the linker
13068 to report an error such as:
13070 @cindex relocation truncated to fit (ColdFire)
13072 relocation truncated to fit: R_68K_GOT16O foobar
13075 If this happens, you should recompile your code with @option{-mxgot}.
13076 It should then work with very large GOTs. However, code generated with
13077 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
13078 the value of a global symbol.
13080 Note that some linkers, including newer versions of the GNU linker,
13081 can create multiple GOTs and sort GOT entries. If you have such a linker,
13082 you should only need to use @option{-mxgot} when compiling a single
13083 object file that accesses more than 8192 GOT entries. Very few do.
13085 These options have no effect unless GCC is generating
13086 position-independent code.
13090 @node M68hc1x Options
13091 @subsection M68hc1x Options
13092 @cindex M68hc1x options
13094 These are the @samp{-m} options defined for the 68hc11 and 68hc12
13095 microcontrollers. The default values for these options depends on
13096 which style of microcontroller was selected when the compiler was configured;
13097 the defaults for the most common choices are given below.
13104 Generate output for a 68HC11. This is the default
13105 when the compiler is configured for 68HC11-based systems.
13111 Generate output for a 68HC12. This is the default
13112 when the compiler is configured for 68HC12-based systems.
13118 Generate output for a 68HCS12.
13120 @item -mauto-incdec
13121 @opindex mauto-incdec
13122 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
13129 Enable the use of 68HC12 min and max instructions.
13132 @itemx -mno-long-calls
13133 @opindex mlong-calls
13134 @opindex mno-long-calls
13135 Treat all calls as being far away (near). If calls are assumed to be
13136 far away, the compiler will use the @code{call} instruction to
13137 call a function and the @code{rtc} instruction for returning.
13141 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13143 @item -msoft-reg-count=@var{count}
13144 @opindex msoft-reg-count
13145 Specify the number of pseudo-soft registers which are used for the
13146 code generation. The maximum number is 32. Using more pseudo-soft
13147 register may or may not result in better code depending on the program.
13148 The default is 4 for 68HC11 and 2 for 68HC12.
13152 @node MCore Options
13153 @subsection MCore Options
13154 @cindex MCore options
13156 These are the @samp{-m} options defined for the Motorola M*Core
13162 @itemx -mno-hardlit
13164 @opindex mno-hardlit
13165 Inline constants into the code stream if it can be done in two
13166 instructions or less.
13172 Use the divide instruction. (Enabled by default).
13174 @item -mrelax-immediate
13175 @itemx -mno-relax-immediate
13176 @opindex mrelax-immediate
13177 @opindex mno-relax-immediate
13178 Allow arbitrary sized immediates in bit operations.
13180 @item -mwide-bitfields
13181 @itemx -mno-wide-bitfields
13182 @opindex mwide-bitfields
13183 @opindex mno-wide-bitfields
13184 Always treat bit-fields as int-sized.
13186 @item -m4byte-functions
13187 @itemx -mno-4byte-functions
13188 @opindex m4byte-functions
13189 @opindex mno-4byte-functions
13190 Force all functions to be aligned to a four byte boundary.
13192 @item -mcallgraph-data
13193 @itemx -mno-callgraph-data
13194 @opindex mcallgraph-data
13195 @opindex mno-callgraph-data
13196 Emit callgraph information.
13199 @itemx -mno-slow-bytes
13200 @opindex mslow-bytes
13201 @opindex mno-slow-bytes
13202 Prefer word access when reading byte quantities.
13204 @item -mlittle-endian
13205 @itemx -mbig-endian
13206 @opindex mlittle-endian
13207 @opindex mbig-endian
13208 Generate code for a little endian target.
13214 Generate code for the 210 processor.
13218 Assume that run-time support has been provided and so omit the
13219 simulator library (@file{libsim.a)} from the linker command line.
13221 @item -mstack-increment=@var{size}
13222 @opindex mstack-increment
13223 Set the maximum amount for a single stack increment operation. Large
13224 values can increase the speed of programs which contain functions
13225 that need a large amount of stack space, but they can also trigger a
13226 segmentation fault if the stack is extended too much. The default
13232 @subsection MeP Options
13233 @cindex MeP options
13239 Enables the @code{abs} instruction, which is the absolute difference
13240 between two registers.
13244 Enables all the optional instructions - average, multiply, divide, bit
13245 operations, leading zero, absolute difference, min/max, clip, and
13251 Enables the @code{ave} instruction, which computes the average of two
13254 @item -mbased=@var{n}
13256 Variables of size @var{n} bytes or smaller will be placed in the
13257 @code{.based} section by default. Based variables use the @code{$tp}
13258 register as a base register, and there is a 128 byte limit to the
13259 @code{.based} section.
13263 Enables the bit operation instructions - bit test (@code{btstm}), set
13264 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
13265 test-and-set (@code{tas}).
13267 @item -mc=@var{name}
13269 Selects which section constant data will be placed in. @var{name} may
13270 be @code{tiny}, @code{near}, or @code{far}.
13274 Enables the @code{clip} instruction. Note that @code{-mclip} is not
13275 useful unless you also provide @code{-mminmax}.
13277 @item -mconfig=@var{name}
13279 Selects one of the build-in core configurations. Each MeP chip has
13280 one or more modules in it; each module has a core CPU and a variety of
13281 coprocessors, optional instructions, and peripherals. The
13282 @code{MeP-Integrator} tool, not part of GCC, provides these
13283 configurations through this option; using this option is the same as
13284 using all the corresponding command line options. The default
13285 configuration is @code{default}.
13289 Enables the coprocessor instructions. By default, this is a 32-bit
13290 coprocessor. Note that the coprocessor is normally enabled via the
13291 @code{-mconfig=} option.
13295 Enables the 32-bit coprocessor's instructions.
13299 Enables the 64-bit coprocessor's instructions.
13303 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
13307 Causes constant variables to be placed in the @code{.near} section.
13311 Enables the @code{div} and @code{divu} instructions.
13315 Generate big-endian code.
13319 Generate little-endian code.
13321 @item -mio-volatile
13322 @opindex mio-volatile
13323 Tells the compiler that any variable marked with the @code{io}
13324 attribute is to be considered volatile.
13328 Causes variables to be assigned to the @code{.far} section by default.
13332 Enables the @code{leadz} (leading zero) instruction.
13336 Causes variables to be assigned to the @code{.near} section by default.
13340 Enables the @code{min} and @code{max} instructions.
13344 Enables the multiplication and multiply-accumulate instructions.
13348 Disables all the optional instructions enabled by @code{-mall-opts}.
13352 Enables the @code{repeat} and @code{erepeat} instructions, used for
13353 low-overhead looping.
13357 Causes all variables to default to the @code{.tiny} section. Note
13358 that there is a 65536 byte limit to this section. Accesses to these
13359 variables use the @code{%gp} base register.
13363 Enables the saturation instructions. Note that the compiler does not
13364 currently generate these itself, but this option is included for
13365 compatibility with other tools, like @code{as}.
13369 Link the SDRAM-based runtime instead of the default ROM-based runtime.
13373 Link the simulator runtime libraries.
13377 Link the simulator runtime libraries, excluding built-in support
13378 for reset and exception vectors and tables.
13382 Causes all functions to default to the @code{.far} section. Without
13383 this option, functions default to the @code{.near} section.
13385 @item -mtiny=@var{n}
13387 Variables that are @var{n} bytes or smaller will be allocated to the
13388 @code{.tiny} section. These variables use the @code{$gp} base
13389 register. The default for this option is 4, but note that there's a
13390 65536 byte limit to the @code{.tiny} section.
13395 @subsection MIPS Options
13396 @cindex MIPS options
13402 Generate big-endian code.
13406 Generate little-endian code. This is the default for @samp{mips*el-*-*}
13409 @item -march=@var{arch}
13411 Generate code that will run on @var{arch}, which can be the name of a
13412 generic MIPS ISA, or the name of a particular processor.
13414 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
13415 @samp{mips32}, @samp{mips32r2}, @samp{mips64} and @samp{mips64r2}.
13416 The processor names are:
13417 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
13418 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
13419 @samp{5kc}, @samp{5kf},
13421 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
13422 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
13423 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
13424 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
13425 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
13426 @samp{loongson2e}, @samp{loongson2f},
13430 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
13431 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
13432 @samp{rm7000}, @samp{rm9000},
13433 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
13436 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
13437 @samp{vr5000}, @samp{vr5400}, @samp{vr5500}
13439 The special value @samp{from-abi} selects the
13440 most compatible architecture for the selected ABI (that is,
13441 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
13443 Native Linux/GNU toolchains also support the value @samp{native},
13444 which selects the best architecture option for the host processor.
13445 @option{-march=native} has no effect if GCC does not recognize
13448 In processor names, a final @samp{000} can be abbreviated as @samp{k}
13449 (for example, @samp{-march=r2k}). Prefixes are optional, and
13450 @samp{vr} may be written @samp{r}.
13452 Names of the form @samp{@var{n}f2_1} refer to processors with
13453 FPUs clocked at half the rate of the core, names of the form
13454 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
13455 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
13456 processors with FPUs clocked a ratio of 3:2 with respect to the core.
13457 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
13458 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
13459 accepted as synonyms for @samp{@var{n}f1_1}.
13461 GCC defines two macros based on the value of this option. The first
13462 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
13463 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
13464 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
13465 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
13466 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
13468 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
13469 above. In other words, it will have the full prefix and will not
13470 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
13471 the macro names the resolved architecture (either @samp{"mips1"} or
13472 @samp{"mips3"}). It names the default architecture when no
13473 @option{-march} option is given.
13475 @item -mtune=@var{arch}
13477 Optimize for @var{arch}. Among other things, this option controls
13478 the way instructions are scheduled, and the perceived cost of arithmetic
13479 operations. The list of @var{arch} values is the same as for
13482 When this option is not used, GCC will optimize for the processor
13483 specified by @option{-march}. By using @option{-march} and
13484 @option{-mtune} together, it is possible to generate code that will
13485 run on a family of processors, but optimize the code for one
13486 particular member of that family.
13488 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
13489 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
13490 @samp{-march} ones described above.
13494 Equivalent to @samp{-march=mips1}.
13498 Equivalent to @samp{-march=mips2}.
13502 Equivalent to @samp{-march=mips3}.
13506 Equivalent to @samp{-march=mips4}.
13510 Equivalent to @samp{-march=mips32}.
13514 Equivalent to @samp{-march=mips32r2}.
13518 Equivalent to @samp{-march=mips64}.
13522 Equivalent to @samp{-march=mips64r2}.
13527 @opindex mno-mips16
13528 Generate (do not generate) MIPS16 code. If GCC is targetting a
13529 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
13531 MIPS16 code generation can also be controlled on a per-function basis
13532 by means of @code{mips16} and @code{nomips16} attributes.
13533 @xref{Function Attributes}, for more information.
13535 @item -mflip-mips16
13536 @opindex mflip-mips16
13537 Generate MIPS16 code on alternating functions. This option is provided
13538 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
13539 not intended for ordinary use in compiling user code.
13541 @item -minterlink-mips16
13542 @itemx -mno-interlink-mips16
13543 @opindex minterlink-mips16
13544 @opindex mno-interlink-mips16
13545 Require (do not require) that non-MIPS16 code be link-compatible with
13548 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
13549 it must either use a call or an indirect jump. @option{-minterlink-mips16}
13550 therefore disables direct jumps unless GCC knows that the target of the
13551 jump is not MIPS16.
13563 Generate code for the given ABI@.
13565 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
13566 generates 64-bit code when you select a 64-bit architecture, but you
13567 can use @option{-mgp32} to get 32-bit code instead.
13569 For information about the O64 ABI, see
13570 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
13572 GCC supports a variant of the o32 ABI in which floating-point registers
13573 are 64 rather than 32 bits wide. You can select this combination with
13574 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
13575 and @samp{mfhc1} instructions and is therefore only supported for
13576 MIPS32R2 processors.
13578 The register assignments for arguments and return values remain the
13579 same, but each scalar value is passed in a single 64-bit register
13580 rather than a pair of 32-bit registers. For example, scalar
13581 floating-point values are returned in @samp{$f0} only, not a
13582 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
13583 remains the same, but all 64 bits are saved.
13586 @itemx -mno-abicalls
13588 @opindex mno-abicalls
13589 Generate (do not generate) code that is suitable for SVR4-style
13590 dynamic objects. @option{-mabicalls} is the default for SVR4-based
13595 Generate (do not generate) code that is fully position-independent,
13596 and that can therefore be linked into shared libraries. This option
13597 only affects @option{-mabicalls}.
13599 All @option{-mabicalls} code has traditionally been position-independent,
13600 regardless of options like @option{-fPIC} and @option{-fpic}. However,
13601 as an extension, the GNU toolchain allows executables to use absolute
13602 accesses for locally-binding symbols. It can also use shorter GP
13603 initialization sequences and generate direct calls to locally-defined
13604 functions. This mode is selected by @option{-mno-shared}.
13606 @option{-mno-shared} depends on binutils 2.16 or higher and generates
13607 objects that can only be linked by the GNU linker. However, the option
13608 does not affect the ABI of the final executable; it only affects the ABI
13609 of relocatable objects. Using @option{-mno-shared} will generally make
13610 executables both smaller and quicker.
13612 @option{-mshared} is the default.
13618 Assume (do not assume) that the static and dynamic linkers
13619 support PLTs and copy relocations. This option only affects
13620 @samp{-mno-shared -mabicalls}. For the n64 ABI, this option
13621 has no effect without @samp{-msym32}.
13623 You can make @option{-mplt} the default by configuring
13624 GCC with @option{--with-mips-plt}. The default is
13625 @option{-mno-plt} otherwise.
13631 Lift (do not lift) the usual restrictions on the size of the global
13634 GCC normally uses a single instruction to load values from the GOT@.
13635 While this is relatively efficient, it will only work if the GOT
13636 is smaller than about 64k. Anything larger will cause the linker
13637 to report an error such as:
13639 @cindex relocation truncated to fit (MIPS)
13641 relocation truncated to fit: R_MIPS_GOT16 foobar
13644 If this happens, you should recompile your code with @option{-mxgot}.
13645 It should then work with very large GOTs, although it will also be
13646 less efficient, since it will take three instructions to fetch the
13647 value of a global symbol.
13649 Note that some linkers can create multiple GOTs. If you have such a
13650 linker, you should only need to use @option{-mxgot} when a single object
13651 file accesses more than 64k's worth of GOT entries. Very few do.
13653 These options have no effect unless GCC is generating position
13658 Assume that general-purpose registers are 32 bits wide.
13662 Assume that general-purpose registers are 64 bits wide.
13666 Assume that floating-point registers are 32 bits wide.
13670 Assume that floating-point registers are 64 bits wide.
13673 @opindex mhard-float
13674 Use floating-point coprocessor instructions.
13677 @opindex msoft-float
13678 Do not use floating-point coprocessor instructions. Implement
13679 floating-point calculations using library calls instead.
13681 @item -msingle-float
13682 @opindex msingle-float
13683 Assume that the floating-point coprocessor only supports single-precision
13686 @item -mdouble-float
13687 @opindex mdouble-float
13688 Assume that the floating-point coprocessor supports double-precision
13689 operations. This is the default.
13695 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
13696 implement atomic memory built-in functions. When neither option is
13697 specified, GCC will use the instructions if the target architecture
13700 @option{-mllsc} is useful if the runtime environment can emulate the
13701 instructions and @option{-mno-llsc} can be useful when compiling for
13702 nonstandard ISAs. You can make either option the default by
13703 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
13704 respectively. @option{--with-llsc} is the default for some
13705 configurations; see the installation documentation for details.
13711 Use (do not use) revision 1 of the MIPS DSP ASE@.
13712 @xref{MIPS DSP Built-in Functions}. This option defines the
13713 preprocessor macro @samp{__mips_dsp}. It also defines
13714 @samp{__mips_dsp_rev} to 1.
13720 Use (do not use) revision 2 of the MIPS DSP ASE@.
13721 @xref{MIPS DSP Built-in Functions}. This option defines the
13722 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
13723 It also defines @samp{__mips_dsp_rev} to 2.
13726 @itemx -mno-smartmips
13727 @opindex msmartmips
13728 @opindex mno-smartmips
13729 Use (do not use) the MIPS SmartMIPS ASE.
13731 @item -mpaired-single
13732 @itemx -mno-paired-single
13733 @opindex mpaired-single
13734 @opindex mno-paired-single
13735 Use (do not use) paired-single floating-point instructions.
13736 @xref{MIPS Paired-Single Support}. This option requires
13737 hardware floating-point support to be enabled.
13743 Use (do not use) MIPS Digital Media Extension instructions.
13744 This option can only be used when generating 64-bit code and requires
13745 hardware floating-point support to be enabled.
13750 @opindex mno-mips3d
13751 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
13752 The option @option{-mips3d} implies @option{-mpaired-single}.
13758 Use (do not use) MT Multithreading instructions.
13762 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
13763 an explanation of the default and the way that the pointer size is
13768 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
13770 The default size of @code{int}s, @code{long}s and pointers depends on
13771 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
13772 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
13773 32-bit @code{long}s. Pointers are the same size as @code{long}s,
13774 or the same size as integer registers, whichever is smaller.
13780 Assume (do not assume) that all symbols have 32-bit values, regardless
13781 of the selected ABI@. This option is useful in combination with
13782 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
13783 to generate shorter and faster references to symbolic addresses.
13787 Put definitions of externally-visible data in a small data section
13788 if that data is no bigger than @var{num} bytes. GCC can then access
13789 the data more efficiently; see @option{-mgpopt} for details.
13791 The default @option{-G} option depends on the configuration.
13793 @item -mlocal-sdata
13794 @itemx -mno-local-sdata
13795 @opindex mlocal-sdata
13796 @opindex mno-local-sdata
13797 Extend (do not extend) the @option{-G} behavior to local data too,
13798 such as to static variables in C@. @option{-mlocal-sdata} is the
13799 default for all configurations.
13801 If the linker complains that an application is using too much small data,
13802 you might want to try rebuilding the less performance-critical parts with
13803 @option{-mno-local-sdata}. You might also want to build large
13804 libraries with @option{-mno-local-sdata}, so that the libraries leave
13805 more room for the main program.
13807 @item -mextern-sdata
13808 @itemx -mno-extern-sdata
13809 @opindex mextern-sdata
13810 @opindex mno-extern-sdata
13811 Assume (do not assume) that externally-defined data will be in
13812 a small data section if that data is within the @option{-G} limit.
13813 @option{-mextern-sdata} is the default for all configurations.
13815 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
13816 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
13817 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
13818 is placed in a small data section. If @var{Var} is defined by another
13819 module, you must either compile that module with a high-enough
13820 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
13821 definition. If @var{Var} is common, you must link the application
13822 with a high-enough @option{-G} setting.
13824 The easiest way of satisfying these restrictions is to compile
13825 and link every module with the same @option{-G} option. However,
13826 you may wish to build a library that supports several different
13827 small data limits. You can do this by compiling the library with
13828 the highest supported @option{-G} setting and additionally using
13829 @option{-mno-extern-sdata} to stop the library from making assumptions
13830 about externally-defined data.
13836 Use (do not use) GP-relative accesses for symbols that are known to be
13837 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
13838 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
13841 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
13842 might not hold the value of @code{_gp}. For example, if the code is
13843 part of a library that might be used in a boot monitor, programs that
13844 call boot monitor routines will pass an unknown value in @code{$gp}.
13845 (In such situations, the boot monitor itself would usually be compiled
13846 with @option{-G0}.)
13848 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
13849 @option{-mno-extern-sdata}.
13851 @item -membedded-data
13852 @itemx -mno-embedded-data
13853 @opindex membedded-data
13854 @opindex mno-embedded-data
13855 Allocate variables to the read-only data section first if possible, then
13856 next in the small data section if possible, otherwise in data. This gives
13857 slightly slower code than the default, but reduces the amount of RAM required
13858 when executing, and thus may be preferred for some embedded systems.
13860 @item -muninit-const-in-rodata
13861 @itemx -mno-uninit-const-in-rodata
13862 @opindex muninit-const-in-rodata
13863 @opindex mno-uninit-const-in-rodata
13864 Put uninitialized @code{const} variables in the read-only data section.
13865 This option is only meaningful in conjunction with @option{-membedded-data}.
13867 @item -mcode-readable=@var{setting}
13868 @opindex mcode-readable
13869 Specify whether GCC may generate code that reads from executable sections.
13870 There are three possible settings:
13873 @item -mcode-readable=yes
13874 Instructions may freely access executable sections. This is the
13877 @item -mcode-readable=pcrel
13878 MIPS16 PC-relative load instructions can access executable sections,
13879 but other instructions must not do so. This option is useful on 4KSc
13880 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
13881 It is also useful on processors that can be configured to have a dual
13882 instruction/data SRAM interface and that, like the M4K, automatically
13883 redirect PC-relative loads to the instruction RAM.
13885 @item -mcode-readable=no
13886 Instructions must not access executable sections. This option can be
13887 useful on targets that are configured to have a dual instruction/data
13888 SRAM interface but that (unlike the M4K) do not automatically redirect
13889 PC-relative loads to the instruction RAM.
13892 @item -msplit-addresses
13893 @itemx -mno-split-addresses
13894 @opindex msplit-addresses
13895 @opindex mno-split-addresses
13896 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
13897 relocation operators. This option has been superseded by
13898 @option{-mexplicit-relocs} but is retained for backwards compatibility.
13900 @item -mexplicit-relocs
13901 @itemx -mno-explicit-relocs
13902 @opindex mexplicit-relocs
13903 @opindex mno-explicit-relocs
13904 Use (do not use) assembler relocation operators when dealing with symbolic
13905 addresses. The alternative, selected by @option{-mno-explicit-relocs},
13906 is to use assembler macros instead.
13908 @option{-mexplicit-relocs} is the default if GCC was configured
13909 to use an assembler that supports relocation operators.
13911 @item -mcheck-zero-division
13912 @itemx -mno-check-zero-division
13913 @opindex mcheck-zero-division
13914 @opindex mno-check-zero-division
13915 Trap (do not trap) on integer division by zero.
13917 The default is @option{-mcheck-zero-division}.
13919 @item -mdivide-traps
13920 @itemx -mdivide-breaks
13921 @opindex mdivide-traps
13922 @opindex mdivide-breaks
13923 MIPS systems check for division by zero by generating either a
13924 conditional trap or a break instruction. Using traps results in
13925 smaller code, but is only supported on MIPS II and later. Also, some
13926 versions of the Linux kernel have a bug that prevents trap from
13927 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
13928 allow conditional traps on architectures that support them and
13929 @option{-mdivide-breaks} to force the use of breaks.
13931 The default is usually @option{-mdivide-traps}, but this can be
13932 overridden at configure time using @option{--with-divide=breaks}.
13933 Divide-by-zero checks can be completely disabled using
13934 @option{-mno-check-zero-division}.
13939 @opindex mno-memcpy
13940 Force (do not force) the use of @code{memcpy()} for non-trivial block
13941 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
13942 most constant-sized copies.
13945 @itemx -mno-long-calls
13946 @opindex mlong-calls
13947 @opindex mno-long-calls
13948 Disable (do not disable) use of the @code{jal} instruction. Calling
13949 functions using @code{jal} is more efficient but requires the caller
13950 and callee to be in the same 256 megabyte segment.
13952 This option has no effect on abicalls code. The default is
13953 @option{-mno-long-calls}.
13959 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
13960 instructions, as provided by the R4650 ISA@.
13963 @itemx -mno-fused-madd
13964 @opindex mfused-madd
13965 @opindex mno-fused-madd
13966 Enable (disable) use of the floating point multiply-accumulate
13967 instructions, when they are available. The default is
13968 @option{-mfused-madd}.
13970 When multiply-accumulate instructions are used, the intermediate
13971 product is calculated to infinite precision and is not subject to
13972 the FCSR Flush to Zero bit. This may be undesirable in some
13977 Tell the MIPS assembler to not run its preprocessor over user
13978 assembler files (with a @samp{.s} suffix) when assembling them.
13981 @itemx -mno-fix-r4000
13982 @opindex mfix-r4000
13983 @opindex mno-fix-r4000
13984 Work around certain R4000 CPU errata:
13987 A double-word or a variable shift may give an incorrect result if executed
13988 immediately after starting an integer division.
13990 A double-word or a variable shift may give an incorrect result if executed
13991 while an integer multiplication is in progress.
13993 An integer division may give an incorrect result if started in a delay slot
13994 of a taken branch or a jump.
13998 @itemx -mno-fix-r4400
13999 @opindex mfix-r4400
14000 @opindex mno-fix-r4400
14001 Work around certain R4400 CPU errata:
14004 A double-word or a variable shift may give an incorrect result if executed
14005 immediately after starting an integer division.
14009 @itemx -mno-fix-r10000
14010 @opindex mfix-r10000
14011 @opindex mno-fix-r10000
14012 Work around certain R10000 errata:
14015 @code{ll}/@code{sc} sequences may not behave atomically on revisions
14016 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
14019 This option can only be used if the target architecture supports
14020 branch-likely instructions. @option{-mfix-r10000} is the default when
14021 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
14025 @itemx -mno-fix-vr4120
14026 @opindex mfix-vr4120
14027 Work around certain VR4120 errata:
14030 @code{dmultu} does not always produce the correct result.
14032 @code{div} and @code{ddiv} do not always produce the correct result if one
14033 of the operands is negative.
14035 The workarounds for the division errata rely on special functions in
14036 @file{libgcc.a}. At present, these functions are only provided by
14037 the @code{mips64vr*-elf} configurations.
14039 Other VR4120 errata require a nop to be inserted between certain pairs of
14040 instructions. These errata are handled by the assembler, not by GCC itself.
14043 @opindex mfix-vr4130
14044 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
14045 workarounds are implemented by the assembler rather than by GCC,
14046 although GCC will avoid using @code{mflo} and @code{mfhi} if the
14047 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
14048 instructions are available instead.
14051 @itemx -mno-fix-sb1
14053 Work around certain SB-1 CPU core errata.
14054 (This flag currently works around the SB-1 revision 2
14055 ``F1'' and ``F2'' floating point errata.)
14057 @item -mr10k-cache-barrier=@var{setting}
14058 @opindex mr10k-cache-barrier
14059 Specify whether GCC should insert cache barriers to avoid the
14060 side-effects of speculation on R10K processors.
14062 In common with many processors, the R10K tries to predict the outcome
14063 of a conditional branch and speculatively executes instructions from
14064 the ``taken'' branch. It later aborts these instructions if the
14065 predicted outcome was wrong. However, on the R10K, even aborted
14066 instructions can have side effects.
14068 This problem only affects kernel stores and, depending on the system,
14069 kernel loads. As an example, a speculatively-executed store may load
14070 the target memory into cache and mark the cache line as dirty, even if
14071 the store itself is later aborted. If a DMA operation writes to the
14072 same area of memory before the ``dirty'' line is flushed, the cached
14073 data will overwrite the DMA-ed data. See the R10K processor manual
14074 for a full description, including other potential problems.
14076 One workaround is to insert cache barrier instructions before every memory
14077 access that might be speculatively executed and that might have side
14078 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
14079 controls GCC's implementation of this workaround. It assumes that
14080 aborted accesses to any byte in the following regions will not have
14085 the memory occupied by the current function's stack frame;
14088 the memory occupied by an incoming stack argument;
14091 the memory occupied by an object with a link-time-constant address.
14094 It is the kernel's responsibility to ensure that speculative
14095 accesses to these regions are indeed safe.
14097 If the input program contains a function declaration such as:
14103 then the implementation of @code{foo} must allow @code{j foo} and
14104 @code{jal foo} to be executed speculatively. GCC honors this
14105 restriction for functions it compiles itself. It expects non-GCC
14106 functions (such as hand-written assembly code) to do the same.
14108 The option has three forms:
14111 @item -mr10k-cache-barrier=load-store
14112 Insert a cache barrier before a load or store that might be
14113 speculatively executed and that might have side effects even
14116 @item -mr10k-cache-barrier=store
14117 Insert a cache barrier before a store that might be speculatively
14118 executed and that might have side effects even if aborted.
14120 @item -mr10k-cache-barrier=none
14121 Disable the insertion of cache barriers. This is the default setting.
14124 @item -mflush-func=@var{func}
14125 @itemx -mno-flush-func
14126 @opindex mflush-func
14127 Specifies the function to call to flush the I and D caches, or to not
14128 call any such function. If called, the function must take the same
14129 arguments as the common @code{_flush_func()}, that is, the address of the
14130 memory range for which the cache is being flushed, the size of the
14131 memory range, and the number 3 (to flush both caches). The default
14132 depends on the target GCC was configured for, but commonly is either
14133 @samp{_flush_func} or @samp{__cpu_flush}.
14135 @item mbranch-cost=@var{num}
14136 @opindex mbranch-cost
14137 Set the cost of branches to roughly @var{num} ``simple'' instructions.
14138 This cost is only a heuristic and is not guaranteed to produce
14139 consistent results across releases. A zero cost redundantly selects
14140 the default, which is based on the @option{-mtune} setting.
14142 @item -mbranch-likely
14143 @itemx -mno-branch-likely
14144 @opindex mbranch-likely
14145 @opindex mno-branch-likely
14146 Enable or disable use of Branch Likely instructions, regardless of the
14147 default for the selected architecture. By default, Branch Likely
14148 instructions may be generated if they are supported by the selected
14149 architecture. An exception is for the MIPS32 and MIPS64 architectures
14150 and processors which implement those architectures; for those, Branch
14151 Likely instructions will not be generated by default because the MIPS32
14152 and MIPS64 architectures specifically deprecate their use.
14154 @item -mfp-exceptions
14155 @itemx -mno-fp-exceptions
14156 @opindex mfp-exceptions
14157 Specifies whether FP exceptions are enabled. This affects how we schedule
14158 FP instructions for some processors. The default is that FP exceptions are
14161 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
14162 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
14165 @item -mvr4130-align
14166 @itemx -mno-vr4130-align
14167 @opindex mvr4130-align
14168 The VR4130 pipeline is two-way superscalar, but can only issue two
14169 instructions together if the first one is 8-byte aligned. When this
14170 option is enabled, GCC will align pairs of instructions that it
14171 thinks should execute in parallel.
14173 This option only has an effect when optimizing for the VR4130.
14174 It normally makes code faster, but at the expense of making it bigger.
14175 It is enabled by default at optimization level @option{-O3}.
14180 Enable (disable) generation of @code{synci} instructions on
14181 architectures that support it. The @code{synci} instructions (if
14182 enabled) will be generated when @code{__builtin___clear_cache()} is
14185 This option defaults to @code{-mno-synci}, but the default can be
14186 overridden by configuring with @code{--with-synci}.
14188 When compiling code for single processor systems, it is generally safe
14189 to use @code{synci}. However, on many multi-core (SMP) systems, it
14190 will not invalidate the instruction caches on all cores and may lead
14191 to undefined behavior.
14193 @item -mrelax-pic-calls
14194 @itemx -mno-relax-pic-calls
14195 @opindex mrelax-pic-calls
14196 Try to turn PIC calls that are normally dispatched via register
14197 @code{$25} into direct calls. This is only possible if the linker can
14198 resolve the destination at link-time and if the destination is within
14199 range for a direct call.
14201 @option{-mrelax-pic-calls} is the default if GCC was configured to use
14202 an assembler and a linker that supports the @code{.reloc} assembly
14203 directive and @code{-mexplicit-relocs} is in effect. With
14204 @code{-mno-explicit-relocs}, this optimization can be performed by the
14205 assembler and the linker alone without help from the compiler.
14209 @subsection MMIX Options
14210 @cindex MMIX Options
14212 These options are defined for the MMIX:
14216 @itemx -mno-libfuncs
14218 @opindex mno-libfuncs
14219 Specify that intrinsic library functions are being compiled, passing all
14220 values in registers, no matter the size.
14223 @itemx -mno-epsilon
14225 @opindex mno-epsilon
14226 Generate floating-point comparison instructions that compare with respect
14227 to the @code{rE} epsilon register.
14229 @item -mabi=mmixware
14231 @opindex mabi=mmixware
14233 Generate code that passes function parameters and return values that (in
14234 the called function) are seen as registers @code{$0} and up, as opposed to
14235 the GNU ABI which uses global registers @code{$231} and up.
14237 @item -mzero-extend
14238 @itemx -mno-zero-extend
14239 @opindex mzero-extend
14240 @opindex mno-zero-extend
14241 When reading data from memory in sizes shorter than 64 bits, use (do not
14242 use) zero-extending load instructions by default, rather than
14243 sign-extending ones.
14246 @itemx -mno-knuthdiv
14248 @opindex mno-knuthdiv
14249 Make the result of a division yielding a remainder have the same sign as
14250 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
14251 remainder follows the sign of the dividend. Both methods are
14252 arithmetically valid, the latter being almost exclusively used.
14254 @item -mtoplevel-symbols
14255 @itemx -mno-toplevel-symbols
14256 @opindex mtoplevel-symbols
14257 @opindex mno-toplevel-symbols
14258 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
14259 code can be used with the @code{PREFIX} assembly directive.
14263 Generate an executable in the ELF format, rather than the default
14264 @samp{mmo} format used by the @command{mmix} simulator.
14266 @item -mbranch-predict
14267 @itemx -mno-branch-predict
14268 @opindex mbranch-predict
14269 @opindex mno-branch-predict
14270 Use (do not use) the probable-branch instructions, when static branch
14271 prediction indicates a probable branch.
14273 @item -mbase-addresses
14274 @itemx -mno-base-addresses
14275 @opindex mbase-addresses
14276 @opindex mno-base-addresses
14277 Generate (do not generate) code that uses @emph{base addresses}. Using a
14278 base address automatically generates a request (handled by the assembler
14279 and the linker) for a constant to be set up in a global register. The
14280 register is used for one or more base address requests within the range 0
14281 to 255 from the value held in the register. The generally leads to short
14282 and fast code, but the number of different data items that can be
14283 addressed is limited. This means that a program that uses lots of static
14284 data may require @option{-mno-base-addresses}.
14286 @item -msingle-exit
14287 @itemx -mno-single-exit
14288 @opindex msingle-exit
14289 @opindex mno-single-exit
14290 Force (do not force) generated code to have a single exit point in each
14294 @node MN10300 Options
14295 @subsection MN10300 Options
14296 @cindex MN10300 options
14298 These @option{-m} options are defined for Matsushita MN10300 architectures:
14303 Generate code to avoid bugs in the multiply instructions for the MN10300
14304 processors. This is the default.
14306 @item -mno-mult-bug
14307 @opindex mno-mult-bug
14308 Do not generate code to avoid bugs in the multiply instructions for the
14309 MN10300 processors.
14313 Generate code which uses features specific to the AM33 processor.
14317 Do not generate code which uses features specific to the AM33 processor. This
14320 @item -mreturn-pointer-on-d0
14321 @opindex mreturn-pointer-on-d0
14322 When generating a function which returns a pointer, return the pointer
14323 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
14324 only in a0, and attempts to call such functions without a prototype
14325 would result in errors. Note that this option is on by default; use
14326 @option{-mno-return-pointer-on-d0} to disable it.
14330 Do not link in the C run-time initialization object file.
14334 Indicate to the linker that it should perform a relaxation optimization pass
14335 to shorten branches, calls and absolute memory addresses. This option only
14336 has an effect when used on the command line for the final link step.
14338 This option makes symbolic debugging impossible.
14341 @node PDP-11 Options
14342 @subsection PDP-11 Options
14343 @cindex PDP-11 Options
14345 These options are defined for the PDP-11:
14350 Use hardware FPP floating point. This is the default. (FIS floating
14351 point on the PDP-11/40 is not supported.)
14354 @opindex msoft-float
14355 Do not use hardware floating point.
14359 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
14363 Return floating-point results in memory. This is the default.
14367 Generate code for a PDP-11/40.
14371 Generate code for a PDP-11/45. This is the default.
14375 Generate code for a PDP-11/10.
14377 @item -mbcopy-builtin
14378 @opindex mbcopy-builtin
14379 Use inline @code{movmemhi} patterns for copying memory. This is the
14384 Do not use inline @code{movmemhi} patterns for copying memory.
14390 Use 16-bit @code{int}. This is the default.
14396 Use 32-bit @code{int}.
14399 @itemx -mno-float32
14401 @opindex mno-float32
14402 Use 64-bit @code{float}. This is the default.
14405 @itemx -mno-float64
14407 @opindex mno-float64
14408 Use 32-bit @code{float}.
14412 Use @code{abshi2} pattern. This is the default.
14416 Do not use @code{abshi2} pattern.
14418 @item -mbranch-expensive
14419 @opindex mbranch-expensive
14420 Pretend that branches are expensive. This is for experimenting with
14421 code generation only.
14423 @item -mbranch-cheap
14424 @opindex mbranch-cheap
14425 Do not pretend that branches are expensive. This is the default.
14429 Generate code for a system with split I&D@.
14433 Generate code for a system without split I&D@. This is the default.
14437 Use Unix assembler syntax. This is the default when configured for
14438 @samp{pdp11-*-bsd}.
14442 Use DEC assembler syntax. This is the default when configured for any
14443 PDP-11 target other than @samp{pdp11-*-bsd}.
14446 @node picoChip Options
14447 @subsection picoChip Options
14448 @cindex picoChip options
14450 These @samp{-m} options are defined for picoChip implementations:
14454 @item -mae=@var{ae_type}
14456 Set the instruction set, register set, and instruction scheduling
14457 parameters for array element type @var{ae_type}. Supported values
14458 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
14460 @option{-mae=ANY} selects a completely generic AE type. Code
14461 generated with this option will run on any of the other AE types. The
14462 code will not be as efficient as it would be if compiled for a specific
14463 AE type, and some types of operation (e.g., multiplication) will not
14464 work properly on all types of AE.
14466 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
14467 for compiled code, and is the default.
14469 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
14470 option may suffer from poor performance of byte (char) manipulation,
14471 since the DSP AE does not provide hardware support for byte load/stores.
14473 @item -msymbol-as-address
14474 Enable the compiler to directly use a symbol name as an address in a
14475 load/store instruction, without first loading it into a
14476 register. Typically, the use of this option will generate larger
14477 programs, which run faster than when the option isn't used. However, the
14478 results vary from program to program, so it is left as a user option,
14479 rather than being permanently enabled.
14481 @item -mno-inefficient-warnings
14482 Disables warnings about the generation of inefficient code. These
14483 warnings can be generated, for example, when compiling code which
14484 performs byte-level memory operations on the MAC AE type. The MAC AE has
14485 no hardware support for byte-level memory operations, so all byte
14486 load/stores must be synthesized from word load/store operations. This is
14487 inefficient and a warning will be generated indicating to the programmer
14488 that they should rewrite the code to avoid byte operations, or to target
14489 an AE type which has the necessary hardware support. This option enables
14490 the warning to be turned off.
14494 @node PowerPC Options
14495 @subsection PowerPC Options
14496 @cindex PowerPC options
14498 These are listed under @xref{RS/6000 and PowerPC Options}.
14500 @node RS/6000 and PowerPC Options
14501 @subsection IBM RS/6000 and PowerPC Options
14502 @cindex RS/6000 and PowerPC Options
14503 @cindex IBM RS/6000 and PowerPC Options
14505 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
14512 @itemx -mno-powerpc
14513 @itemx -mpowerpc-gpopt
14514 @itemx -mno-powerpc-gpopt
14515 @itemx -mpowerpc-gfxopt
14516 @itemx -mno-powerpc-gfxopt
14518 @itemx -mno-powerpc64
14522 @itemx -mno-popcntb
14524 @itemx -mno-popcntd
14532 @itemx -mno-hard-dfp
14536 @opindex mno-power2
14538 @opindex mno-powerpc
14539 @opindex mpowerpc-gpopt
14540 @opindex mno-powerpc-gpopt
14541 @opindex mpowerpc-gfxopt
14542 @opindex mno-powerpc-gfxopt
14543 @opindex mpowerpc64
14544 @opindex mno-powerpc64
14548 @opindex mno-popcntb
14550 @opindex mno-popcntd
14556 @opindex mno-mfpgpr
14558 @opindex mno-hard-dfp
14559 GCC supports two related instruction set architectures for the
14560 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
14561 instructions supported by the @samp{rios} chip set used in the original
14562 RS/6000 systems and the @dfn{PowerPC} instruction set is the
14563 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
14564 the IBM 4xx, 6xx, and follow-on microprocessors.
14566 Neither architecture is a subset of the other. However there is a
14567 large common subset of instructions supported by both. An MQ
14568 register is included in processors supporting the POWER architecture.
14570 You use these options to specify which instructions are available on the
14571 processor you are using. The default value of these options is
14572 determined when configuring GCC@. Specifying the
14573 @option{-mcpu=@var{cpu_type}} overrides the specification of these
14574 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
14575 rather than the options listed above.
14577 The @option{-mpower} option allows GCC to generate instructions that
14578 are found only in the POWER architecture and to use the MQ register.
14579 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
14580 to generate instructions that are present in the POWER2 architecture but
14581 not the original POWER architecture.
14583 The @option{-mpowerpc} option allows GCC to generate instructions that
14584 are found only in the 32-bit subset of the PowerPC architecture.
14585 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
14586 GCC to use the optional PowerPC architecture instructions in the
14587 General Purpose group, including floating-point square root. Specifying
14588 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
14589 use the optional PowerPC architecture instructions in the Graphics
14590 group, including floating-point select.
14592 The @option{-mmfcrf} option allows GCC to generate the move from
14593 condition register field instruction implemented on the POWER4
14594 processor and other processors that support the PowerPC V2.01
14596 The @option{-mpopcntb} option allows GCC to generate the popcount and
14597 double precision FP reciprocal estimate instruction implemented on the
14598 POWER5 processor and other processors that support the PowerPC V2.02
14600 The @option{-mpopcntd} option allows GCC to generate the popcount
14601 instruction implemented on the POWER7 processor and other processors
14602 that support the PowerPC V2.06 architecture.
14603 The @option{-mfprnd} option allows GCC to generate the FP round to
14604 integer instructions implemented on the POWER5+ processor and other
14605 processors that support the PowerPC V2.03 architecture.
14606 The @option{-mcmpb} option allows GCC to generate the compare bytes
14607 instruction implemented on the POWER6 processor and other processors
14608 that support the PowerPC V2.05 architecture.
14609 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
14610 general purpose register instructions implemented on the POWER6X
14611 processor and other processors that support the extended PowerPC V2.05
14613 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
14614 point instructions implemented on some POWER processors.
14616 The @option{-mpowerpc64} option allows GCC to generate the additional
14617 64-bit instructions that are found in the full PowerPC64 architecture
14618 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
14619 @option{-mno-powerpc64}.
14621 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
14622 will use only the instructions in the common subset of both
14623 architectures plus some special AIX common-mode calls, and will not use
14624 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
14625 permits GCC to use any instruction from either architecture and to
14626 allow use of the MQ register; specify this for the Motorola MPC601.
14628 @item -mnew-mnemonics
14629 @itemx -mold-mnemonics
14630 @opindex mnew-mnemonics
14631 @opindex mold-mnemonics
14632 Select which mnemonics to use in the generated assembler code. With
14633 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
14634 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
14635 assembler mnemonics defined for the POWER architecture. Instructions
14636 defined in only one architecture have only one mnemonic; GCC uses that
14637 mnemonic irrespective of which of these options is specified.
14639 GCC defaults to the mnemonics appropriate for the architecture in
14640 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
14641 value of these option. Unless you are building a cross-compiler, you
14642 should normally not specify either @option{-mnew-mnemonics} or
14643 @option{-mold-mnemonics}, but should instead accept the default.
14645 @item -mcpu=@var{cpu_type}
14647 Set architecture type, register usage, choice of mnemonics, and
14648 instruction scheduling parameters for machine type @var{cpu_type}.
14649 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
14650 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
14651 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
14652 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
14653 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
14654 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
14655 @samp{e300c3}, @samp{e500mc}, @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5},
14656 @samp{power}, @samp{power2}, @samp{power3}, @samp{power4},
14657 @samp{power5}, @samp{power5+}, @samp{power6}, @samp{power6x}, @samp{power7},
14658 @samp{common}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
14659 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
14661 @option{-mcpu=common} selects a completely generic processor. Code
14662 generated under this option will run on any POWER or PowerPC processor.
14663 GCC will use only the instructions in the common subset of both
14664 architectures, and will not use the MQ register. GCC assumes a generic
14665 processor model for scheduling purposes.
14667 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
14668 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
14669 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
14670 types, with an appropriate, generic processor model assumed for
14671 scheduling purposes.
14673 The other options specify a specific processor. Code generated under
14674 those options will run best on that processor, and may not run at all on
14677 The @option{-mcpu} options automatically enable or disable the
14680 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
14681 -mnew-mnemonics -mpopcntb -mpopcntd -mpower -mpower2 -mpowerpc64 @gol
14682 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
14683 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx}
14685 The particular options set for any particular CPU will vary between
14686 compiler versions, depending on what setting seems to produce optimal
14687 code for that CPU; it doesn't necessarily reflect the actual hardware's
14688 capabilities. If you wish to set an individual option to a particular
14689 value, you may specify it after the @option{-mcpu} option, like
14690 @samp{-mcpu=970 -mno-altivec}.
14692 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
14693 not enabled or disabled by the @option{-mcpu} option at present because
14694 AIX does not have full support for these options. You may still
14695 enable or disable them individually if you're sure it'll work in your
14698 @item -mtune=@var{cpu_type}
14700 Set the instruction scheduling parameters for machine type
14701 @var{cpu_type}, but do not set the architecture type, register usage, or
14702 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
14703 values for @var{cpu_type} are used for @option{-mtune} as for
14704 @option{-mcpu}. If both are specified, the code generated will use the
14705 architecture, registers, and mnemonics set by @option{-mcpu}, but the
14706 scheduling parameters set by @option{-mtune}.
14712 Generate code to compute division as reciprocal estimate and iterative
14713 refinement, creating opportunities for increased throughput. This
14714 feature requires: optional PowerPC Graphics instruction set for single
14715 precision and FRE instruction for double precision, assuming divides
14716 cannot generate user-visible traps, and the domain values not include
14717 Infinities, denormals or zero denominator.
14720 @itemx -mno-altivec
14722 @opindex mno-altivec
14723 Generate code that uses (does not use) AltiVec instructions, and also
14724 enable the use of built-in functions that allow more direct access to
14725 the AltiVec instruction set. You may also need to set
14726 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
14732 @opindex mno-vrsave
14733 Generate VRSAVE instructions when generating AltiVec code.
14735 @item -mgen-cell-microcode
14736 @opindex mgen-cell-microcode
14737 Generate Cell microcode instructions
14739 @item -mwarn-cell-microcode
14740 @opindex mwarn-cell-microcode
14741 Warning when a Cell microcode instruction is going to emitted. An example
14742 of a Cell microcode instruction is a variable shift.
14745 @opindex msecure-plt
14746 Generate code that allows ld and ld.so to build executables and shared
14747 libraries with non-exec .plt and .got sections. This is a PowerPC
14748 32-bit SYSV ABI option.
14752 Generate code that uses a BSS .plt section that ld.so fills in, and
14753 requires .plt and .got sections that are both writable and executable.
14754 This is a PowerPC 32-bit SYSV ABI option.
14760 This switch enables or disables the generation of ISEL instructions.
14762 @item -misel=@var{yes/no}
14763 This switch has been deprecated. Use @option{-misel} and
14764 @option{-mno-isel} instead.
14770 This switch enables or disables the generation of SPE simd
14776 @opindex mno-paired
14777 This switch enables or disables the generation of PAIRED simd
14780 @item -mspe=@var{yes/no}
14781 This option has been deprecated. Use @option{-mspe} and
14782 @option{-mno-spe} instead.
14788 Generate code that uses (does not use) vector/scalar (VSX)
14789 instructions, and also enable the use of built-in functions that allow
14790 more direct access to the VSX instruction set.
14792 @item -mfloat-gprs=@var{yes/single/double/no}
14793 @itemx -mfloat-gprs
14794 @opindex mfloat-gprs
14795 This switch enables or disables the generation of floating point
14796 operations on the general purpose registers for architectures that
14799 The argument @var{yes} or @var{single} enables the use of
14800 single-precision floating point operations.
14802 The argument @var{double} enables the use of single and
14803 double-precision floating point operations.
14805 The argument @var{no} disables floating point operations on the
14806 general purpose registers.
14808 This option is currently only available on the MPC854x.
14814 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
14815 targets (including GNU/Linux). The 32-bit environment sets int, long
14816 and pointer to 32 bits and generates code that runs on any PowerPC
14817 variant. The 64-bit environment sets int to 32 bits and long and
14818 pointer to 64 bits, and generates code for PowerPC64, as for
14819 @option{-mpowerpc64}.
14822 @itemx -mno-fp-in-toc
14823 @itemx -mno-sum-in-toc
14824 @itemx -mminimal-toc
14826 @opindex mno-fp-in-toc
14827 @opindex mno-sum-in-toc
14828 @opindex mminimal-toc
14829 Modify generation of the TOC (Table Of Contents), which is created for
14830 every executable file. The @option{-mfull-toc} option is selected by
14831 default. In that case, GCC will allocate at least one TOC entry for
14832 each unique non-automatic variable reference in your program. GCC
14833 will also place floating-point constants in the TOC@. However, only
14834 16,384 entries are available in the TOC@.
14836 If you receive a linker error message that saying you have overflowed
14837 the available TOC space, you can reduce the amount of TOC space used
14838 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
14839 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
14840 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
14841 generate code to calculate the sum of an address and a constant at
14842 run-time instead of putting that sum into the TOC@. You may specify one
14843 or both of these options. Each causes GCC to produce very slightly
14844 slower and larger code at the expense of conserving TOC space.
14846 If you still run out of space in the TOC even when you specify both of
14847 these options, specify @option{-mminimal-toc} instead. This option causes
14848 GCC to make only one TOC entry for every file. When you specify this
14849 option, GCC will produce code that is slower and larger but which
14850 uses extremely little TOC space. You may wish to use this option
14851 only on files that contain less frequently executed code.
14857 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
14858 @code{long} type, and the infrastructure needed to support them.
14859 Specifying @option{-maix64} implies @option{-mpowerpc64} and
14860 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
14861 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
14864 @itemx -mno-xl-compat
14865 @opindex mxl-compat
14866 @opindex mno-xl-compat
14867 Produce code that conforms more closely to IBM XL compiler semantics
14868 when using AIX-compatible ABI@. Pass floating-point arguments to
14869 prototyped functions beyond the register save area (RSA) on the stack
14870 in addition to argument FPRs. Do not assume that most significant
14871 double in 128-bit long double value is properly rounded when comparing
14872 values and converting to double. Use XL symbol names for long double
14875 The AIX calling convention was extended but not initially documented to
14876 handle an obscure K&R C case of calling a function that takes the
14877 address of its arguments with fewer arguments than declared. IBM XL
14878 compilers access floating point arguments which do not fit in the
14879 RSA from the stack when a subroutine is compiled without
14880 optimization. Because always storing floating-point arguments on the
14881 stack is inefficient and rarely needed, this option is not enabled by
14882 default and only is necessary when calling subroutines compiled by IBM
14883 XL compilers without optimization.
14887 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
14888 application written to use message passing with special startup code to
14889 enable the application to run. The system must have PE installed in the
14890 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
14891 must be overridden with the @option{-specs=} option to specify the
14892 appropriate directory location. The Parallel Environment does not
14893 support threads, so the @option{-mpe} option and the @option{-pthread}
14894 option are incompatible.
14896 @item -malign-natural
14897 @itemx -malign-power
14898 @opindex malign-natural
14899 @opindex malign-power
14900 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
14901 @option{-malign-natural} overrides the ABI-defined alignment of larger
14902 types, such as floating-point doubles, on their natural size-based boundary.
14903 The option @option{-malign-power} instructs GCC to follow the ABI-specified
14904 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
14906 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
14910 @itemx -mhard-float
14911 @opindex msoft-float
14912 @opindex mhard-float
14913 Generate code that does not use (uses) the floating-point register set.
14914 Software floating point emulation is provided if you use the
14915 @option{-msoft-float} option, and pass the option to GCC when linking.
14917 @item -msingle-float
14918 @itemx -mdouble-float
14919 @opindex msingle-float
14920 @opindex mdouble-float
14921 Generate code for single or double-precision floating point operations.
14922 @option{-mdouble-float} implies @option{-msingle-float}.
14925 @opindex msimple-fpu
14926 Do not generate sqrt and div instructions for hardware floating point unit.
14930 Specify type of floating point unit. Valid values are @var{sp_lite}
14931 (equivalent to -msingle-float -msimple-fpu), @var{dp_lite} (equivalent
14932 to -mdouble-float -msimple-fpu), @var{sp_full} (equivalent to -msingle-float),
14933 and @var{dp_full} (equivalent to -mdouble-float).
14936 @opindex mxilinx-fpu
14937 Perform optimizations for floating point unit on Xilinx PPC 405/440.
14940 @itemx -mno-multiple
14942 @opindex mno-multiple
14943 Generate code that uses (does not use) the load multiple word
14944 instructions and the store multiple word instructions. These
14945 instructions are generated by default on POWER systems, and not
14946 generated on PowerPC systems. Do not use @option{-mmultiple} on little
14947 endian PowerPC systems, since those instructions do not work when the
14948 processor is in little endian mode. The exceptions are PPC740 and
14949 PPC750 which permit the instructions usage in little endian mode.
14954 @opindex mno-string
14955 Generate code that uses (does not use) the load string instructions
14956 and the store string word instructions to save multiple registers and
14957 do small block moves. These instructions are generated by default on
14958 POWER systems, and not generated on PowerPC systems. Do not use
14959 @option{-mstring} on little endian PowerPC systems, since those
14960 instructions do not work when the processor is in little endian mode.
14961 The exceptions are PPC740 and PPC750 which permit the instructions
14962 usage in little endian mode.
14967 @opindex mno-update
14968 Generate code that uses (does not use) the load or store instructions
14969 that update the base register to the address of the calculated memory
14970 location. These instructions are generated by default. If you use
14971 @option{-mno-update}, there is a small window between the time that the
14972 stack pointer is updated and the address of the previous frame is
14973 stored, which means code that walks the stack frame across interrupts or
14974 signals may get corrupted data.
14976 @item -mavoid-indexed-addresses
14977 @item -mno-avoid-indexed-addresses
14978 @opindex mavoid-indexed-addresses
14979 @opindex mno-avoid-indexed-addresses
14980 Generate code that tries to avoid (not avoid) the use of indexed load
14981 or store instructions. These instructions can incur a performance
14982 penalty on Power6 processors in certain situations, such as when
14983 stepping through large arrays that cross a 16M boundary. This option
14984 is enabled by default when targetting Power6 and disabled otherwise.
14987 @itemx -mno-fused-madd
14988 @opindex mfused-madd
14989 @opindex mno-fused-madd
14990 Generate code that uses (does not use) the floating point multiply and
14991 accumulate instructions. These instructions are generated by default if
14992 hardware floating is used.
14998 Generate code that uses (does not use) the half-word multiply and
14999 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
15000 These instructions are generated by default when targetting those
15007 Generate code that uses (does not use) the string-search @samp{dlmzb}
15008 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
15009 generated by default when targetting those processors.
15011 @item -mno-bit-align
15013 @opindex mno-bit-align
15014 @opindex mbit-align
15015 On System V.4 and embedded PowerPC systems do not (do) force structures
15016 and unions that contain bit-fields to be aligned to the base type of the
15019 For example, by default a structure containing nothing but 8
15020 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
15021 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
15022 the structure would be aligned to a 1 byte boundary and be one byte in
15025 @item -mno-strict-align
15026 @itemx -mstrict-align
15027 @opindex mno-strict-align
15028 @opindex mstrict-align
15029 On System V.4 and embedded PowerPC systems do not (do) assume that
15030 unaligned memory references will be handled by the system.
15032 @item -mrelocatable
15033 @itemx -mno-relocatable
15034 @opindex mrelocatable
15035 @opindex mno-relocatable
15036 On embedded PowerPC systems generate code that allows (does not allow)
15037 the program to be relocated to a different address at runtime. If you
15038 use @option{-mrelocatable} on any module, all objects linked together must
15039 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
15041 @item -mrelocatable-lib
15042 @itemx -mno-relocatable-lib
15043 @opindex mrelocatable-lib
15044 @opindex mno-relocatable-lib
15045 On embedded PowerPC systems generate code that allows (does not allow)
15046 the program to be relocated to a different address at runtime. Modules
15047 compiled with @option{-mrelocatable-lib} can be linked with either modules
15048 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
15049 with modules compiled with the @option{-mrelocatable} options.
15055 On System V.4 and embedded PowerPC systems do not (do) assume that
15056 register 2 contains a pointer to a global area pointing to the addresses
15057 used in the program.
15060 @itemx -mlittle-endian
15062 @opindex mlittle-endian
15063 On System V.4 and embedded PowerPC systems compile code for the
15064 processor in little endian mode. The @option{-mlittle-endian} option is
15065 the same as @option{-mlittle}.
15068 @itemx -mbig-endian
15070 @opindex mbig-endian
15071 On System V.4 and embedded PowerPC systems compile code for the
15072 processor in big endian mode. The @option{-mbig-endian} option is
15073 the same as @option{-mbig}.
15075 @item -mdynamic-no-pic
15076 @opindex mdynamic-no-pic
15077 On Darwin and Mac OS X systems, compile code so that it is not
15078 relocatable, but that its external references are relocatable. The
15079 resulting code is suitable for applications, but not shared
15082 @item -mprioritize-restricted-insns=@var{priority}
15083 @opindex mprioritize-restricted-insns
15084 This option controls the priority that is assigned to
15085 dispatch-slot restricted instructions during the second scheduling
15086 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
15087 @var{no/highest/second-highest} priority to dispatch slot restricted
15090 @item -msched-costly-dep=@var{dependence_type}
15091 @opindex msched-costly-dep
15092 This option controls which dependences are considered costly
15093 by the target during instruction scheduling. The argument
15094 @var{dependence_type} takes one of the following values:
15095 @var{no}: no dependence is costly,
15096 @var{all}: all dependences are costly,
15097 @var{true_store_to_load}: a true dependence from store to load is costly,
15098 @var{store_to_load}: any dependence from store to load is costly,
15099 @var{number}: any dependence which latency >= @var{number} is costly.
15101 @item -minsert-sched-nops=@var{scheme}
15102 @opindex minsert-sched-nops
15103 This option controls which nop insertion scheme will be used during
15104 the second scheduling pass. The argument @var{scheme} takes one of the
15106 @var{no}: Don't insert nops.
15107 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
15108 according to the scheduler's grouping.
15109 @var{regroup_exact}: Insert nops to force costly dependent insns into
15110 separate groups. Insert exactly as many nops as needed to force an insn
15111 to a new group, according to the estimated processor grouping.
15112 @var{number}: Insert nops to force costly dependent insns into
15113 separate groups. Insert @var{number} nops to force an insn to a new group.
15116 @opindex mcall-sysv
15117 On System V.4 and embedded PowerPC systems compile code using calling
15118 conventions that adheres to the March 1995 draft of the System V
15119 Application Binary Interface, PowerPC processor supplement. This is the
15120 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
15122 @item -mcall-sysv-eabi
15124 @opindex mcall-sysv-eabi
15125 @opindex mcall-eabi
15126 Specify both @option{-mcall-sysv} and @option{-meabi} options.
15128 @item -mcall-sysv-noeabi
15129 @opindex mcall-sysv-noeabi
15130 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
15132 @item -mcall-aixdesc
15134 On System V.4 and embedded PowerPC systems compile code for the AIX
15138 @opindex mcall-linux
15139 On System V.4 and embedded PowerPC systems compile code for the
15140 Linux-based GNU system.
15144 On System V.4 and embedded PowerPC systems compile code for the
15145 Hurd-based GNU system.
15147 @item -mcall-freebsd
15148 @opindex mcall-freebsd
15149 On System V.4 and embedded PowerPC systems compile code for the
15150 FreeBSD operating system.
15152 @item -mcall-netbsd
15153 @opindex mcall-netbsd
15154 On System V.4 and embedded PowerPC systems compile code for the
15155 NetBSD operating system.
15157 @item -mcall-openbsd
15158 @opindex mcall-netbsd
15159 On System V.4 and embedded PowerPC systems compile code for the
15160 OpenBSD operating system.
15162 @item -maix-struct-return
15163 @opindex maix-struct-return
15164 Return all structures in memory (as specified by the AIX ABI)@.
15166 @item -msvr4-struct-return
15167 @opindex msvr4-struct-return
15168 Return structures smaller than 8 bytes in registers (as specified by the
15171 @item -mabi=@var{abi-type}
15173 Extend the current ABI with a particular extension, or remove such extension.
15174 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
15175 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
15179 Extend the current ABI with SPE ABI extensions. This does not change
15180 the default ABI, instead it adds the SPE ABI extensions to the current
15184 @opindex mabi=no-spe
15185 Disable Booke SPE ABI extensions for the current ABI@.
15187 @item -mabi=ibmlongdouble
15188 @opindex mabi=ibmlongdouble
15189 Change the current ABI to use IBM extended precision long double.
15190 This is a PowerPC 32-bit SYSV ABI option.
15192 @item -mabi=ieeelongdouble
15193 @opindex mabi=ieeelongdouble
15194 Change the current ABI to use IEEE extended precision long double.
15195 This is a PowerPC 32-bit Linux ABI option.
15198 @itemx -mno-prototype
15199 @opindex mprototype
15200 @opindex mno-prototype
15201 On System V.4 and embedded PowerPC systems assume that all calls to
15202 variable argument functions are properly prototyped. Otherwise, the
15203 compiler must insert an instruction before every non prototyped call to
15204 set or clear bit 6 of the condition code register (@var{CR}) to
15205 indicate whether floating point values were passed in the floating point
15206 registers in case the function takes a variable arguments. With
15207 @option{-mprototype}, only calls to prototyped variable argument functions
15208 will set or clear the bit.
15212 On embedded PowerPC systems, assume that the startup module is called
15213 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
15214 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
15219 On embedded PowerPC systems, assume that the startup module is called
15220 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
15225 On embedded PowerPC systems, assume that the startup module is called
15226 @file{crt0.o} and the standard C libraries are @file{libads.a} and
15229 @item -myellowknife
15230 @opindex myellowknife
15231 On embedded PowerPC systems, assume that the startup module is called
15232 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
15237 On System V.4 and embedded PowerPC systems, specify that you are
15238 compiling for a VxWorks system.
15242 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
15243 header to indicate that @samp{eabi} extended relocations are used.
15249 On System V.4 and embedded PowerPC systems do (do not) adhere to the
15250 Embedded Applications Binary Interface (eabi) which is a set of
15251 modifications to the System V.4 specifications. Selecting @option{-meabi}
15252 means that the stack is aligned to an 8 byte boundary, a function
15253 @code{__eabi} is called to from @code{main} to set up the eabi
15254 environment, and the @option{-msdata} option can use both @code{r2} and
15255 @code{r13} to point to two separate small data areas. Selecting
15256 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
15257 do not call an initialization function from @code{main}, and the
15258 @option{-msdata} option will only use @code{r13} to point to a single
15259 small data area. The @option{-meabi} option is on by default if you
15260 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
15263 @opindex msdata=eabi
15264 On System V.4 and embedded PowerPC systems, put small initialized
15265 @code{const} global and static data in the @samp{.sdata2} section, which
15266 is pointed to by register @code{r2}. Put small initialized
15267 non-@code{const} global and static data in the @samp{.sdata} section,
15268 which is pointed to by register @code{r13}. Put small uninitialized
15269 global and static data in the @samp{.sbss} section, which is adjacent to
15270 the @samp{.sdata} section. The @option{-msdata=eabi} option is
15271 incompatible with the @option{-mrelocatable} option. The
15272 @option{-msdata=eabi} option also sets the @option{-memb} option.
15275 @opindex msdata=sysv
15276 On System V.4 and embedded PowerPC systems, put small global and static
15277 data in the @samp{.sdata} section, which is pointed to by register
15278 @code{r13}. Put small uninitialized global and static data in the
15279 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
15280 The @option{-msdata=sysv} option is incompatible with the
15281 @option{-mrelocatable} option.
15283 @item -msdata=default
15285 @opindex msdata=default
15287 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
15288 compile code the same as @option{-msdata=eabi}, otherwise compile code the
15289 same as @option{-msdata=sysv}.
15292 @opindex msdata=data
15293 On System V.4 and embedded PowerPC systems, put small global
15294 data in the @samp{.sdata} section. Put small uninitialized global
15295 data in the @samp{.sbss} section. Do not use register @code{r13}
15296 to address small data however. This is the default behavior unless
15297 other @option{-msdata} options are used.
15301 @opindex msdata=none
15303 On embedded PowerPC systems, put all initialized global and static data
15304 in the @samp{.data} section, and all uninitialized data in the
15305 @samp{.bss} section.
15309 @cindex smaller data references (PowerPC)
15310 @cindex .sdata/.sdata2 references (PowerPC)
15311 On embedded PowerPC systems, put global and static items less than or
15312 equal to @var{num} bytes into the small data or bss sections instead of
15313 the normal data or bss section. By default, @var{num} is 8. The
15314 @option{-G @var{num}} switch is also passed to the linker.
15315 All modules should be compiled with the same @option{-G @var{num}} value.
15318 @itemx -mno-regnames
15320 @opindex mno-regnames
15321 On System V.4 and embedded PowerPC systems do (do not) emit register
15322 names in the assembly language output using symbolic forms.
15325 @itemx -mno-longcall
15327 @opindex mno-longcall
15328 By default assume that all calls are far away so that a longer more
15329 expensive calling sequence is required. This is required for calls
15330 further than 32 megabytes (33,554,432 bytes) from the current location.
15331 A short call will be generated if the compiler knows
15332 the call cannot be that far away. This setting can be overridden by
15333 the @code{shortcall} function attribute, or by @code{#pragma
15336 Some linkers are capable of detecting out-of-range calls and generating
15337 glue code on the fly. On these systems, long calls are unnecessary and
15338 generate slower code. As of this writing, the AIX linker can do this,
15339 as can the GNU linker for PowerPC/64. It is planned to add this feature
15340 to the GNU linker for 32-bit PowerPC systems as well.
15342 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
15343 callee, L42'', plus a ``branch island'' (glue code). The two target
15344 addresses represent the callee and the ``branch island''. The
15345 Darwin/PPC linker will prefer the first address and generate a ``bl
15346 callee'' if the PPC ``bl'' instruction will reach the callee directly;
15347 otherwise, the linker will generate ``bl L42'' to call the ``branch
15348 island''. The ``branch island'' is appended to the body of the
15349 calling function; it computes the full 32-bit address of the callee
15352 On Mach-O (Darwin) systems, this option directs the compiler emit to
15353 the glue for every direct call, and the Darwin linker decides whether
15354 to use or discard it.
15356 In the future, we may cause GCC to ignore all longcall specifications
15357 when the linker is known to generate glue.
15359 @item -mtls-markers
15360 @itemx -mno-tls-markers
15361 @opindex mtls-markers
15362 @opindex mno-tls-markers
15363 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
15364 specifying the function argument. The relocation allows ld to
15365 reliably associate function call with argument setup instructions for
15366 TLS optimization, which in turn allows gcc to better schedule the
15371 Adds support for multithreading with the @dfn{pthreads} library.
15372 This option sets flags for both the preprocessor and linker.
15377 @subsection RX Options
15380 These @option{-m} options are defined for RX implementations:
15383 @item -m64bit-doubles
15384 @itemx -m32bit-doubles
15385 @opindex m64bit-doubles
15386 @opindex m32bit-doubles
15387 Make the @code{double} data type be 64-bits (@option{-m64bit-doubles})
15388 or 32-bits (@option{-m32bit-doubles}) in size. The default is
15389 @option{-m32bit-doubles}. @emph{Note} the RX's hardware floating
15390 point instructions are only used for 32-bit floating point values, and
15391 then only if @option{-ffast-math} has been specified on the command
15392 line. This is because the RX FPU instructions do not properly support
15393 denormal (or sub-normal) values.
15395 @item -mbig-endian-data
15396 @itemx -mlittle-endian-data
15397 @opindex mbig-endian-data
15398 @opindex mlittle-endian-data
15399 Store data (but not code) in the big-endian format. The default is
15400 @option{-mlittle-endian-data}, ie to store data in the little endian
15403 @item -msmall-data-limit=@var{N}
15404 @opindex msmall-data-limit
15405 Specifies the maximum size in bytes of global and static variables
15406 which can be placed into the small data area. Using the small data
15407 area can lead to smaller and faster code, but the size of area is
15408 limited and it is up to the programmer to ensure that the area does
15409 not overflow. Also when the small data area is used one of the RX's
15410 registers (@code{r13}) is reserved for use pointing to this area, so
15411 it is no longer available for use by the compiler. This could result
15412 in slower and/or larger code if variables which once could have been
15413 held in @code{r13} are now pushed onto the stack.
15415 Note, common variables (variables which have not been initialised) and
15416 constants are not placed into the small data area as they are assigned
15417 to other sections in the output executeable.
15419 The default value is zero, which disables this feature. Note, this
15420 feature is not enabled by default with higher optimization levels
15421 (@option{-O2} etc) because of the potentially deterimental effects of
15422 reserving register @code{r13}. It is up to the programmer to
15423 experiment and discover whether this feature is of benefit to their
15430 Use the simulator runtime. The default is to use the libgloss board
15433 @item -mas100-syntax
15434 @item -mno-as100-syntax
15435 @opindex mas100-syntax
15436 @opindex mno-as100-syntax
15437 When generating assembler output use a syntax that is compatible with
15438 Renesas's AS100 assembler. This syntax can also be handled by the GAS
15439 assembler but it has some restrictions so generating it is not the
15442 @item -mmax-constant-size=@var{N}
15443 @opindex mmax-constant-size
15444 Specifies the maxium size, in bytes, of a constant that can be used as
15445 an operand in a RX instruction. Although the RX instruction set does
15446 allow consants of up to 4 bytes in length to be used in instructions,
15447 a longer value equates to a longer instruction. Thus in some
15448 circumstances it can be beneficial to restrict the size of constants
15449 that are used in instructions. Constants that are too big are instead
15450 placed into a constant pool and referenced via register indirection.
15452 The value @var{N} can be between 0 and 3. A value of 0, the default,
15453 means that constants of any size are allowed.
15457 Enable linker relaxation. Linker relaxation is a process whereby the
15458 linker will attempt to reduce the size of a program by finding shorter
15459 versions of various instructions. Disabled by default.
15461 @item -mint-register=@var{N}
15462 @opindex mint-register
15463 Specify the number of registers to reserve for fast interrupt handler
15464 functions. The value @var{N} can be between 0 and 4. A value of 1
15465 means that register @code{r13} will be reserved for ther exclusive use
15466 of fast interrupt handlers. A value of 2 reserves @code{r13} and
15467 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
15468 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
15469 A value of 0, the default, does not reserve any registers.
15472 @emph{Note:} The generic GCC command line @option{-ffixed-@var{reg}}
15473 has special significance to the RX port when used with the
15474 @code{interrupt} function attribute. This attribute indicates a
15475 function intended to process fast interrupts. GCC will will ensure
15476 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
15477 and/or @code{r13} and only provided that the normal use of the
15478 corresponding registers have been restricted via the
15479 @option{-ffixed-@var{reg}} or @option{-mint-register} command line
15482 @node S/390 and zSeries Options
15483 @subsection S/390 and zSeries Options
15484 @cindex S/390 and zSeries Options
15486 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
15490 @itemx -msoft-float
15491 @opindex mhard-float
15492 @opindex msoft-float
15493 Use (do not use) the hardware floating-point instructions and registers
15494 for floating-point operations. When @option{-msoft-float} is specified,
15495 functions in @file{libgcc.a} will be used to perform floating-point
15496 operations. When @option{-mhard-float} is specified, the compiler
15497 generates IEEE floating-point instructions. This is the default.
15500 @itemx -mno-hard-dfp
15502 @opindex mno-hard-dfp
15503 Use (do not use) the hardware decimal-floating-point instructions for
15504 decimal-floating-point operations. When @option{-mno-hard-dfp} is
15505 specified, functions in @file{libgcc.a} will be used to perform
15506 decimal-floating-point operations. When @option{-mhard-dfp} is
15507 specified, the compiler generates decimal-floating-point hardware
15508 instructions. This is the default for @option{-march=z9-ec} or higher.
15510 @item -mlong-double-64
15511 @itemx -mlong-double-128
15512 @opindex mlong-double-64
15513 @opindex mlong-double-128
15514 These switches control the size of @code{long double} type. A size
15515 of 64bit makes the @code{long double} type equivalent to the @code{double}
15516 type. This is the default.
15519 @itemx -mno-backchain
15520 @opindex mbackchain
15521 @opindex mno-backchain
15522 Store (do not store) the address of the caller's frame as backchain pointer
15523 into the callee's stack frame.
15524 A backchain may be needed to allow debugging using tools that do not understand
15525 DWARF-2 call frame information.
15526 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
15527 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
15528 the backchain is placed into the topmost word of the 96/160 byte register
15531 In general, code compiled with @option{-mbackchain} is call-compatible with
15532 code compiled with @option{-mmo-backchain}; however, use of the backchain
15533 for debugging purposes usually requires that the whole binary is built with
15534 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
15535 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15536 to build a linux kernel use @option{-msoft-float}.
15538 The default is to not maintain the backchain.
15540 @item -mpacked-stack
15541 @itemx -mno-packed-stack
15542 @opindex mpacked-stack
15543 @opindex mno-packed-stack
15544 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
15545 specified, the compiler uses the all fields of the 96/160 byte register save
15546 area only for their default purpose; unused fields still take up stack space.
15547 When @option{-mpacked-stack} is specified, register save slots are densely
15548 packed at the top of the register save area; unused space is reused for other
15549 purposes, allowing for more efficient use of the available stack space.
15550 However, when @option{-mbackchain} is also in effect, the topmost word of
15551 the save area is always used to store the backchain, and the return address
15552 register is always saved two words below the backchain.
15554 As long as the stack frame backchain is not used, code generated with
15555 @option{-mpacked-stack} is call-compatible with code generated with
15556 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
15557 S/390 or zSeries generated code that uses the stack frame backchain at run
15558 time, not just for debugging purposes. Such code is not call-compatible
15559 with code compiled with @option{-mpacked-stack}. Also, note that the
15560 combination of @option{-mbackchain},
15561 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15562 to build a linux kernel use @option{-msoft-float}.
15564 The default is to not use the packed stack layout.
15567 @itemx -mno-small-exec
15568 @opindex msmall-exec
15569 @opindex mno-small-exec
15570 Generate (or do not generate) code using the @code{bras} instruction
15571 to do subroutine calls.
15572 This only works reliably if the total executable size does not
15573 exceed 64k. The default is to use the @code{basr} instruction instead,
15574 which does not have this limitation.
15580 When @option{-m31} is specified, generate code compliant to the
15581 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
15582 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
15583 particular to generate 64-bit instructions. For the @samp{s390}
15584 targets, the default is @option{-m31}, while the @samp{s390x}
15585 targets default to @option{-m64}.
15591 When @option{-mzarch} is specified, generate code using the
15592 instructions available on z/Architecture.
15593 When @option{-mesa} is specified, generate code using the
15594 instructions available on ESA/390. Note that @option{-mesa} is
15595 not possible with @option{-m64}.
15596 When generating code compliant to the GNU/Linux for S/390 ABI,
15597 the default is @option{-mesa}. When generating code compliant
15598 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
15604 Generate (or do not generate) code using the @code{mvcle} instruction
15605 to perform block moves. When @option{-mno-mvcle} is specified,
15606 use a @code{mvc} loop instead. This is the default unless optimizing for
15613 Print (or do not print) additional debug information when compiling.
15614 The default is to not print debug information.
15616 @item -march=@var{cpu-type}
15618 Generate code that will run on @var{cpu-type}, which is the name of a system
15619 representing a certain processor type. Possible values for
15620 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, @samp{z990},
15621 @samp{z9-109}, @samp{z9-ec} and @samp{z10}.
15622 When generating code using the instructions available on z/Architecture,
15623 the default is @option{-march=z900}. Otherwise, the default is
15624 @option{-march=g5}.
15626 @item -mtune=@var{cpu-type}
15628 Tune to @var{cpu-type} everything applicable about the generated code,
15629 except for the ABI and the set of available instructions.
15630 The list of @var{cpu-type} values is the same as for @option{-march}.
15631 The default is the value used for @option{-march}.
15634 @itemx -mno-tpf-trace
15635 @opindex mtpf-trace
15636 @opindex mno-tpf-trace
15637 Generate code that adds (does not add) in TPF OS specific branches to trace
15638 routines in the operating system. This option is off by default, even
15639 when compiling for the TPF OS@.
15642 @itemx -mno-fused-madd
15643 @opindex mfused-madd
15644 @opindex mno-fused-madd
15645 Generate code that uses (does not use) the floating point multiply and
15646 accumulate instructions. These instructions are generated by default if
15647 hardware floating point is used.
15649 @item -mwarn-framesize=@var{framesize}
15650 @opindex mwarn-framesize
15651 Emit a warning if the current function exceeds the given frame size. Because
15652 this is a compile time check it doesn't need to be a real problem when the program
15653 runs. It is intended to identify functions which most probably cause
15654 a stack overflow. It is useful to be used in an environment with limited stack
15655 size e.g.@: the linux kernel.
15657 @item -mwarn-dynamicstack
15658 @opindex mwarn-dynamicstack
15659 Emit a warning if the function calls alloca or uses dynamically
15660 sized arrays. This is generally a bad idea with a limited stack size.
15662 @item -mstack-guard=@var{stack-guard}
15663 @itemx -mstack-size=@var{stack-size}
15664 @opindex mstack-guard
15665 @opindex mstack-size
15666 If these options are provided the s390 back end emits additional instructions in
15667 the function prologue which trigger a trap if the stack size is @var{stack-guard}
15668 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
15669 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
15670 the frame size of the compiled function is chosen.
15671 These options are intended to be used to help debugging stack overflow problems.
15672 The additionally emitted code causes only little overhead and hence can also be
15673 used in production like systems without greater performance degradation. The given
15674 values have to be exact powers of 2 and @var{stack-size} has to be greater than
15675 @var{stack-guard} without exceeding 64k.
15676 In order to be efficient the extra code makes the assumption that the stack starts
15677 at an address aligned to the value given by @var{stack-size}.
15678 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
15681 @node Score Options
15682 @subsection Score Options
15683 @cindex Score Options
15685 These options are defined for Score implementations:
15690 Compile code for big endian mode. This is the default.
15694 Compile code for little endian mode.
15698 Disable generate bcnz instruction.
15702 Enable generate unaligned load and store instruction.
15706 Enable the use of multiply-accumulate instructions. Disabled by default.
15710 Specify the SCORE5 as the target architecture.
15714 Specify the SCORE5U of the target architecture.
15718 Specify the SCORE7 as the target architecture. This is the default.
15722 Specify the SCORE7D as the target architecture.
15726 @subsection SH Options
15728 These @samp{-m} options are defined for the SH implementations:
15733 Generate code for the SH1.
15737 Generate code for the SH2.
15740 Generate code for the SH2e.
15744 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
15745 that the floating-point unit is not used.
15747 @item -m2a-single-only
15748 @opindex m2a-single-only
15749 Generate code for the SH2a-FPU, in such a way that no double-precision
15750 floating point operations are used.
15753 @opindex m2a-single
15754 Generate code for the SH2a-FPU assuming the floating-point unit is in
15755 single-precision mode by default.
15759 Generate code for the SH2a-FPU assuming the floating-point unit is in
15760 double-precision mode by default.
15764 Generate code for the SH3.
15768 Generate code for the SH3e.
15772 Generate code for the SH4 without a floating-point unit.
15774 @item -m4-single-only
15775 @opindex m4-single-only
15776 Generate code for the SH4 with a floating-point unit that only
15777 supports single-precision arithmetic.
15781 Generate code for the SH4 assuming the floating-point unit is in
15782 single-precision mode by default.
15786 Generate code for the SH4.
15790 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
15791 floating-point unit is not used.
15793 @item -m4a-single-only
15794 @opindex m4a-single-only
15795 Generate code for the SH4a, in such a way that no double-precision
15796 floating point operations are used.
15799 @opindex m4a-single
15800 Generate code for the SH4a assuming the floating-point unit is in
15801 single-precision mode by default.
15805 Generate code for the SH4a.
15809 Same as @option{-m4a-nofpu}, except that it implicitly passes
15810 @option{-dsp} to the assembler. GCC doesn't generate any DSP
15811 instructions at the moment.
15815 Compile code for the processor in big endian mode.
15819 Compile code for the processor in little endian mode.
15823 Align doubles at 64-bit boundaries. Note that this changes the calling
15824 conventions, and thus some functions from the standard C library will
15825 not work unless you recompile it first with @option{-mdalign}.
15829 Shorten some address references at link time, when possible; uses the
15830 linker option @option{-relax}.
15834 Use 32-bit offsets in @code{switch} tables. The default is to use
15839 Enable the use of bit manipulation instructions on SH2A.
15843 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
15844 alignment constraints.
15848 Comply with the calling conventions defined by Renesas.
15852 Comply with the calling conventions defined by Renesas.
15856 Comply with the calling conventions defined for GCC before the Renesas
15857 conventions were available. This option is the default for all
15858 targets of the SH toolchain except for @samp{sh-symbianelf}.
15861 @opindex mnomacsave
15862 Mark the @code{MAC} register as call-clobbered, even if
15863 @option{-mhitachi} is given.
15867 Increase IEEE-compliance of floating-point code.
15868 At the moment, this is equivalent to @option{-fno-finite-math-only}.
15869 When generating 16 bit SH opcodes, getting IEEE-conforming results for
15870 comparisons of NANs / infinities incurs extra overhead in every
15871 floating point comparison, therefore the default is set to
15872 @option{-ffinite-math-only}.
15874 @item -minline-ic_invalidate
15875 @opindex minline-ic_invalidate
15876 Inline code to invalidate instruction cache entries after setting up
15877 nested function trampolines.
15878 This option has no effect if -musermode is in effect and the selected
15879 code generation option (e.g. -m4) does not allow the use of the icbi
15881 If the selected code generation option does not allow the use of the icbi
15882 instruction, and -musermode is not in effect, the inlined code will
15883 manipulate the instruction cache address array directly with an associative
15884 write. This not only requires privileged mode, but it will also
15885 fail if the cache line had been mapped via the TLB and has become unmapped.
15889 Dump instruction size and location in the assembly code.
15892 @opindex mpadstruct
15893 This option is deprecated. It pads structures to multiple of 4 bytes,
15894 which is incompatible with the SH ABI@.
15898 Optimize for space instead of speed. Implied by @option{-Os}.
15901 @opindex mprefergot
15902 When generating position-independent code, emit function calls using
15903 the Global Offset Table instead of the Procedure Linkage Table.
15907 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
15908 if the inlined code would not work in user mode.
15909 This is the default when the target is @code{sh-*-linux*}.
15911 @item -multcost=@var{number}
15912 @opindex multcost=@var{number}
15913 Set the cost to assume for a multiply insn.
15915 @item -mdiv=@var{strategy}
15916 @opindex mdiv=@var{strategy}
15917 Set the division strategy to use for SHmedia code. @var{strategy} must be
15918 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
15919 inv:call2, inv:fp .
15920 "fp" performs the operation in floating point. This has a very high latency,
15921 but needs only a few instructions, so it might be a good choice if
15922 your code has enough easily exploitable ILP to allow the compiler to
15923 schedule the floating point instructions together with other instructions.
15924 Division by zero causes a floating point exception.
15925 "inv" uses integer operations to calculate the inverse of the divisor,
15926 and then multiplies the dividend with the inverse. This strategy allows
15927 cse and hoisting of the inverse calculation. Division by zero calculates
15928 an unspecified result, but does not trap.
15929 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
15930 have been found, or if the entire operation has been hoisted to the same
15931 place, the last stages of the inverse calculation are intertwined with the
15932 final multiply to reduce the overall latency, at the expense of using a few
15933 more instructions, and thus offering fewer scheduling opportunities with
15935 "call" calls a library function that usually implements the inv:minlat
15937 This gives high code density for m5-*media-nofpu compilations.
15938 "call2" uses a different entry point of the same library function, where it
15939 assumes that a pointer to a lookup table has already been set up, which
15940 exposes the pointer load to cse / code hoisting optimizations.
15941 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
15942 code generation, but if the code stays unoptimized, revert to the "call",
15943 "call2", or "fp" strategies, respectively. Note that the
15944 potentially-trapping side effect of division by zero is carried by a
15945 separate instruction, so it is possible that all the integer instructions
15946 are hoisted out, but the marker for the side effect stays where it is.
15947 A recombination to fp operations or a call is not possible in that case.
15948 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
15949 that the inverse calculation was nor separated from the multiply, they speed
15950 up division where the dividend fits into 20 bits (plus sign where applicable),
15951 by inserting a test to skip a number of operations in this case; this test
15952 slows down the case of larger dividends. inv20u assumes the case of a such
15953 a small dividend to be unlikely, and inv20l assumes it to be likely.
15955 @item -mdivsi3_libfunc=@var{name}
15956 @opindex mdivsi3_libfunc=@var{name}
15957 Set the name of the library function used for 32 bit signed division to
15958 @var{name}. This only affect the name used in the call and inv:call
15959 division strategies, and the compiler will still expect the same
15960 sets of input/output/clobbered registers as if this option was not present.
15962 @item -mfixed-range=@var{register-range}
15963 @opindex mfixed-range
15964 Generate code treating the given register range as fixed registers.
15965 A fixed register is one that the register allocator can not use. This is
15966 useful when compiling kernel code. A register range is specified as
15967 two registers separated by a dash. Multiple register ranges can be
15968 specified separated by a comma.
15970 @item -madjust-unroll
15971 @opindex madjust-unroll
15972 Throttle unrolling to avoid thrashing target registers.
15973 This option only has an effect if the gcc code base supports the
15974 TARGET_ADJUST_UNROLL_MAX target hook.
15976 @item -mindexed-addressing
15977 @opindex mindexed-addressing
15978 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
15979 This is only safe if the hardware and/or OS implement 32 bit wrap-around
15980 semantics for the indexed addressing mode. The architecture allows the
15981 implementation of processors with 64 bit MMU, which the OS could use to
15982 get 32 bit addressing, but since no current hardware implementation supports
15983 this or any other way to make the indexed addressing mode safe to use in
15984 the 32 bit ABI, the default is -mno-indexed-addressing.
15986 @item -mgettrcost=@var{number}
15987 @opindex mgettrcost=@var{number}
15988 Set the cost assumed for the gettr instruction to @var{number}.
15989 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
15993 Assume pt* instructions won't trap. This will generally generate better
15994 scheduled code, but is unsafe on current hardware. The current architecture
15995 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
15996 This has the unintentional effect of making it unsafe to schedule ptabs /
15997 ptrel before a branch, or hoist it out of a loop. For example,
15998 __do_global_ctors, a part of libgcc that runs constructors at program
15999 startup, calls functions in a list which is delimited by @minus{}1. With the
16000 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
16001 That means that all the constructors will be run a bit quicker, but when
16002 the loop comes to the end of the list, the program crashes because ptabs
16003 loads @minus{}1 into a target register. Since this option is unsafe for any
16004 hardware implementing the current architecture specification, the default
16005 is -mno-pt-fixed. Unless the user specifies a specific cost with
16006 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
16007 this deters register allocation using target registers for storing
16010 @item -minvalid-symbols
16011 @opindex minvalid-symbols
16012 Assume symbols might be invalid. Ordinary function symbols generated by
16013 the compiler will always be valid to load with movi/shori/ptabs or
16014 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
16015 to generate symbols that will cause ptabs / ptrel to trap.
16016 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
16017 It will then prevent cross-basic-block cse, hoisting and most scheduling
16018 of symbol loads. The default is @option{-mno-invalid-symbols}.
16021 @node SPARC Options
16022 @subsection SPARC Options
16023 @cindex SPARC options
16025 These @samp{-m} options are supported on the SPARC:
16028 @item -mno-app-regs
16030 @opindex mno-app-regs
16032 Specify @option{-mapp-regs} to generate output using the global registers
16033 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
16036 To be fully SVR4 ABI compliant at the cost of some performance loss,
16037 specify @option{-mno-app-regs}. You should compile libraries and system
16038 software with this option.
16041 @itemx -mhard-float
16043 @opindex mhard-float
16044 Generate output containing floating point instructions. This is the
16048 @itemx -msoft-float
16050 @opindex msoft-float
16051 Generate output containing library calls for floating point.
16052 @strong{Warning:} the requisite libraries are not available for all SPARC
16053 targets. Normally the facilities of the machine's usual C compiler are
16054 used, but this cannot be done directly in cross-compilation. You must make
16055 your own arrangements to provide suitable library functions for
16056 cross-compilation. The embedded targets @samp{sparc-*-aout} and
16057 @samp{sparclite-*-*} do provide software floating point support.
16059 @option{-msoft-float} changes the calling convention in the output file;
16060 therefore, it is only useful if you compile @emph{all} of a program with
16061 this option. In particular, you need to compile @file{libgcc.a}, the
16062 library that comes with GCC, with @option{-msoft-float} in order for
16065 @item -mhard-quad-float
16066 @opindex mhard-quad-float
16067 Generate output containing quad-word (long double) floating point
16070 @item -msoft-quad-float
16071 @opindex msoft-quad-float
16072 Generate output containing library calls for quad-word (long double)
16073 floating point instructions. The functions called are those specified
16074 in the SPARC ABI@. This is the default.
16076 As of this writing, there are no SPARC implementations that have hardware
16077 support for the quad-word floating point instructions. They all invoke
16078 a trap handler for one of these instructions, and then the trap handler
16079 emulates the effect of the instruction. Because of the trap handler overhead,
16080 this is much slower than calling the ABI library routines. Thus the
16081 @option{-msoft-quad-float} option is the default.
16083 @item -mno-unaligned-doubles
16084 @itemx -munaligned-doubles
16085 @opindex mno-unaligned-doubles
16086 @opindex munaligned-doubles
16087 Assume that doubles have 8 byte alignment. This is the default.
16089 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
16090 alignment only if they are contained in another type, or if they have an
16091 absolute address. Otherwise, it assumes they have 4 byte alignment.
16092 Specifying this option avoids some rare compatibility problems with code
16093 generated by other compilers. It is not the default because it results
16094 in a performance loss, especially for floating point code.
16096 @item -mno-faster-structs
16097 @itemx -mfaster-structs
16098 @opindex mno-faster-structs
16099 @opindex mfaster-structs
16100 With @option{-mfaster-structs}, the compiler assumes that structures
16101 should have 8 byte alignment. This enables the use of pairs of
16102 @code{ldd} and @code{std} instructions for copies in structure
16103 assignment, in place of twice as many @code{ld} and @code{st} pairs.
16104 However, the use of this changed alignment directly violates the SPARC
16105 ABI@. Thus, it's intended only for use on targets where the developer
16106 acknowledges that their resulting code will not be directly in line with
16107 the rules of the ABI@.
16109 @item -mimpure-text
16110 @opindex mimpure-text
16111 @option{-mimpure-text}, used in addition to @option{-shared}, tells
16112 the compiler to not pass @option{-z text} to the linker when linking a
16113 shared object. Using this option, you can link position-dependent
16114 code into a shared object.
16116 @option{-mimpure-text} suppresses the ``relocations remain against
16117 allocatable but non-writable sections'' linker error message.
16118 However, the necessary relocations will trigger copy-on-write, and the
16119 shared object is not actually shared across processes. Instead of
16120 using @option{-mimpure-text}, you should compile all source code with
16121 @option{-fpic} or @option{-fPIC}.
16123 This option is only available on SunOS and Solaris.
16125 @item -mcpu=@var{cpu_type}
16127 Set the instruction set, register set, and instruction scheduling parameters
16128 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
16129 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
16130 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
16131 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
16132 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
16134 Default instruction scheduling parameters are used for values that select
16135 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
16136 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
16138 Here is a list of each supported architecture and their supported
16143 v8: supersparc, hypersparc
16144 sparclite: f930, f934, sparclite86x
16146 v9: ultrasparc, ultrasparc3, niagara, niagara2
16149 By default (unless configured otherwise), GCC generates code for the V7
16150 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
16151 additionally optimizes it for the Cypress CY7C602 chip, as used in the
16152 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
16153 SPARCStation 1, 2, IPX etc.
16155 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
16156 architecture. The only difference from V7 code is that the compiler emits
16157 the integer multiply and integer divide instructions which exist in SPARC-V8
16158 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
16159 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
16162 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
16163 the SPARC architecture. This adds the integer multiply, integer divide step
16164 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
16165 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
16166 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
16167 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
16168 MB86934 chip, which is the more recent SPARClite with FPU@.
16170 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
16171 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
16172 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
16173 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
16174 optimizes it for the TEMIC SPARClet chip.
16176 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
16177 architecture. This adds 64-bit integer and floating-point move instructions,
16178 3 additional floating-point condition code registers and conditional move
16179 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
16180 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
16181 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
16182 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
16183 @option{-mcpu=niagara}, the compiler additionally optimizes it for
16184 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
16185 additionally optimizes it for Sun UltraSPARC T2 chips.
16187 @item -mtune=@var{cpu_type}
16189 Set the instruction scheduling parameters for machine type
16190 @var{cpu_type}, but do not set the instruction set or register set that the
16191 option @option{-mcpu=@var{cpu_type}} would.
16193 The same values for @option{-mcpu=@var{cpu_type}} can be used for
16194 @option{-mtune=@var{cpu_type}}, but the only useful values are those
16195 that select a particular cpu implementation. Those are @samp{cypress},
16196 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
16197 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
16198 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
16203 @opindex mno-v8plus
16204 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
16205 difference from the V8 ABI is that the global and out registers are
16206 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
16207 mode for all SPARC-V9 processors.
16213 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
16214 Visual Instruction Set extensions. The default is @option{-mno-vis}.
16217 These @samp{-m} options are supported in addition to the above
16218 on SPARC-V9 processors in 64-bit environments:
16221 @item -mlittle-endian
16222 @opindex mlittle-endian
16223 Generate code for a processor running in little-endian mode. It is only
16224 available for a few configurations and most notably not on Solaris and Linux.
16230 Generate code for a 32-bit or 64-bit environment.
16231 The 32-bit environment sets int, long and pointer to 32 bits.
16232 The 64-bit environment sets int to 32 bits and long and pointer
16235 @item -mcmodel=medlow
16236 @opindex mcmodel=medlow
16237 Generate code for the Medium/Low code model: 64-bit addresses, programs
16238 must be linked in the low 32 bits of memory. Programs can be statically
16239 or dynamically linked.
16241 @item -mcmodel=medmid
16242 @opindex mcmodel=medmid
16243 Generate code for the Medium/Middle code model: 64-bit addresses, programs
16244 must be linked in the low 44 bits of memory, the text and data segments must
16245 be less than 2GB in size and the data segment must be located within 2GB of
16248 @item -mcmodel=medany
16249 @opindex mcmodel=medany
16250 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
16251 may be linked anywhere in memory, the text and data segments must be less
16252 than 2GB in size and the data segment must be located within 2GB of the
16255 @item -mcmodel=embmedany
16256 @opindex mcmodel=embmedany
16257 Generate code for the Medium/Anywhere code model for embedded systems:
16258 64-bit addresses, the text and data segments must be less than 2GB in
16259 size, both starting anywhere in memory (determined at link time). The
16260 global register %g4 points to the base of the data segment. Programs
16261 are statically linked and PIC is not supported.
16264 @itemx -mno-stack-bias
16265 @opindex mstack-bias
16266 @opindex mno-stack-bias
16267 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
16268 frame pointer if present, are offset by @minus{}2047 which must be added back
16269 when making stack frame references. This is the default in 64-bit mode.
16270 Otherwise, assume no such offset is present.
16273 These switches are supported in addition to the above on Solaris:
16278 Add support for multithreading using the Solaris threads library. This
16279 option sets flags for both the preprocessor and linker. This option does
16280 not affect the thread safety of object code produced by the compiler or
16281 that of libraries supplied with it.
16285 Add support for multithreading using the POSIX threads library. This
16286 option sets flags for both the preprocessor and linker. This option does
16287 not affect the thread safety of object code produced by the compiler or
16288 that of libraries supplied with it.
16292 This is a synonym for @option{-pthreads}.
16296 @subsection SPU Options
16297 @cindex SPU options
16299 These @samp{-m} options are supported on the SPU:
16303 @itemx -merror-reloc
16304 @opindex mwarn-reloc
16305 @opindex merror-reloc
16307 The loader for SPU does not handle dynamic relocations. By default, GCC
16308 will give an error when it generates code that requires a dynamic
16309 relocation. @option{-mno-error-reloc} disables the error,
16310 @option{-mwarn-reloc} will generate a warning instead.
16313 @itemx -munsafe-dma
16315 @opindex munsafe-dma
16317 Instructions which initiate or test completion of DMA must not be
16318 reordered with respect to loads and stores of the memory which is being
16319 accessed. Users typically address this problem using the volatile
16320 keyword, but that can lead to inefficient code in places where the
16321 memory is known to not change. Rather than mark the memory as volatile
16322 we treat the DMA instructions as potentially effecting all memory. With
16323 @option{-munsafe-dma} users must use the volatile keyword to protect
16326 @item -mbranch-hints
16327 @opindex mbranch-hints
16329 By default, GCC will generate a branch hint instruction to avoid
16330 pipeline stalls for always taken or probably taken branches. A hint
16331 will not be generated closer than 8 instructions away from its branch.
16332 There is little reason to disable them, except for debugging purposes,
16333 or to make an object a little bit smaller.
16337 @opindex msmall-mem
16338 @opindex mlarge-mem
16340 By default, GCC generates code assuming that addresses are never larger
16341 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
16342 a full 32 bit address.
16347 By default, GCC links against startup code that assumes the SPU-style
16348 main function interface (which has an unconventional parameter list).
16349 With @option{-mstdmain}, GCC will link your program against startup
16350 code that assumes a C99-style interface to @code{main}, including a
16351 local copy of @code{argv} strings.
16353 @item -mfixed-range=@var{register-range}
16354 @opindex mfixed-range
16355 Generate code treating the given register range as fixed registers.
16356 A fixed register is one that the register allocator can not use. This is
16357 useful when compiling kernel code. A register range is specified as
16358 two registers separated by a dash. Multiple register ranges can be
16359 specified separated by a comma.
16362 @itemx -mdual-nops=@var{n}
16363 @opindex mdual-nops
16364 By default, GCC will insert nops to increase dual issue when it expects
16365 it to increase performance. @var{n} can be a value from 0 to 10. A
16366 smaller @var{n} will insert fewer nops. 10 is the default, 0 is the
16367 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
16369 @item -mhint-max-nops=@var{n}
16370 @opindex mhint-max-nops
16371 Maximum number of nops to insert for a branch hint. A branch hint must
16372 be at least 8 instructions away from the branch it is effecting. GCC
16373 will insert up to @var{n} nops to enforce this, otherwise it will not
16374 generate the branch hint.
16376 @item -mhint-max-distance=@var{n}
16377 @opindex mhint-max-distance
16378 The encoding of the branch hint instruction limits the hint to be within
16379 256 instructions of the branch it is effecting. By default, GCC makes
16380 sure it is within 125.
16383 @opindex msafe-hints
16384 Work around a hardware bug which causes the SPU to stall indefinitely.
16385 By default, GCC will insert the @code{hbrp} instruction to make sure
16386 this stall won't happen.
16390 @node System V Options
16391 @subsection Options for System V
16393 These additional options are available on System V Release 4 for
16394 compatibility with other compilers on those systems:
16399 Create a shared object.
16400 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
16404 Identify the versions of each tool used by the compiler, in a
16405 @code{.ident} assembler directive in the output.
16409 Refrain from adding @code{.ident} directives to the output file (this is
16412 @item -YP,@var{dirs}
16414 Search the directories @var{dirs}, and no others, for libraries
16415 specified with @option{-l}.
16417 @item -Ym,@var{dir}
16419 Look in the directory @var{dir} to find the M4 preprocessor.
16420 The assembler uses this option.
16421 @c This is supposed to go with a -Yd for predefined M4 macro files, but
16422 @c the generic assembler that comes with Solaris takes just -Ym.
16426 @subsection V850 Options
16427 @cindex V850 Options
16429 These @samp{-m} options are defined for V850 implementations:
16433 @itemx -mno-long-calls
16434 @opindex mlong-calls
16435 @opindex mno-long-calls
16436 Treat all calls as being far away (near). If calls are assumed to be
16437 far away, the compiler will always load the functions address up into a
16438 register, and call indirect through the pointer.
16444 Do not optimize (do optimize) basic blocks that use the same index
16445 pointer 4 or more times to copy pointer into the @code{ep} register, and
16446 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
16447 option is on by default if you optimize.
16449 @item -mno-prolog-function
16450 @itemx -mprolog-function
16451 @opindex mno-prolog-function
16452 @opindex mprolog-function
16453 Do not use (do use) external functions to save and restore registers
16454 at the prologue and epilogue of a function. The external functions
16455 are slower, but use less code space if more than one function saves
16456 the same number of registers. The @option{-mprolog-function} option
16457 is on by default if you optimize.
16461 Try to make the code as small as possible. At present, this just turns
16462 on the @option{-mep} and @option{-mprolog-function} options.
16464 @item -mtda=@var{n}
16466 Put static or global variables whose size is @var{n} bytes or less into
16467 the tiny data area that register @code{ep} points to. The tiny data
16468 area can hold up to 256 bytes in total (128 bytes for byte references).
16470 @item -msda=@var{n}
16472 Put static or global variables whose size is @var{n} bytes or less into
16473 the small data area that register @code{gp} points to. The small data
16474 area can hold up to 64 kilobytes.
16476 @item -mzda=@var{n}
16478 Put static or global variables whose size is @var{n} bytes or less into
16479 the first 32 kilobytes of memory.
16483 Specify that the target processor is the V850.
16486 @opindex mbig-switch
16487 Generate code suitable for big switch tables. Use this option only if
16488 the assembler/linker complain about out of range branches within a switch
16493 This option will cause r2 and r5 to be used in the code generated by
16494 the compiler. This setting is the default.
16496 @item -mno-app-regs
16497 @opindex mno-app-regs
16498 This option will cause r2 and r5 to be treated as fixed registers.
16502 Specify that the target processor is the V850E1. The preprocessor
16503 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
16504 this option is used.
16508 Specify that the target processor is the V850E@. The preprocessor
16509 constant @samp{__v850e__} will be defined if this option is used.
16511 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
16512 are defined then a default target processor will be chosen and the
16513 relevant @samp{__v850*__} preprocessor constant will be defined.
16515 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
16516 defined, regardless of which processor variant is the target.
16518 @item -mdisable-callt
16519 @opindex mdisable-callt
16520 This option will suppress generation of the CALLT instruction for the
16521 v850e and v850e1 flavors of the v850 architecture. The default is
16522 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
16527 @subsection VAX Options
16528 @cindex VAX options
16530 These @samp{-m} options are defined for the VAX:
16535 Do not output certain jump instructions (@code{aobleq} and so on)
16536 that the Unix assembler for the VAX cannot handle across long
16541 Do output those jump instructions, on the assumption that you
16542 will assemble with the GNU assembler.
16546 Output code for g-format floating point numbers instead of d-format.
16549 @node VxWorks Options
16550 @subsection VxWorks Options
16551 @cindex VxWorks Options
16553 The options in this section are defined for all VxWorks targets.
16554 Options specific to the target hardware are listed with the other
16555 options for that target.
16560 GCC can generate code for both VxWorks kernels and real time processes
16561 (RTPs). This option switches from the former to the latter. It also
16562 defines the preprocessor macro @code{__RTP__}.
16565 @opindex non-static
16566 Link an RTP executable against shared libraries rather than static
16567 libraries. The options @option{-static} and @option{-shared} can
16568 also be used for RTPs (@pxref{Link Options}); @option{-static}
16575 These options are passed down to the linker. They are defined for
16576 compatibility with Diab.
16579 @opindex Xbind-lazy
16580 Enable lazy binding of function calls. This option is equivalent to
16581 @option{-Wl,-z,now} and is defined for compatibility with Diab.
16585 Disable lazy binding of function calls. This option is the default and
16586 is defined for compatibility with Diab.
16589 @node x86-64 Options
16590 @subsection x86-64 Options
16591 @cindex x86-64 options
16593 These are listed under @xref{i386 and x86-64 Options}.
16595 @node i386 and x86-64 Windows Options
16596 @subsection i386 and x86-64 Windows Options
16597 @cindex i386 and x86-64 Windows Options
16599 These additional options are available for Windows targets:
16604 This option is available for Cygwin and MinGW targets. It
16605 specifies that a console application is to be generated, by
16606 instructing the linker to set the PE header subsystem type
16607 required for console applications.
16608 This is the default behavior for Cygwin and MinGW targets.
16612 This option is available for Cygwin targets. It specifies that
16613 the Cygwin internal interface is to be used for predefined
16614 preprocessor macros, C runtime libraries and related linker
16615 paths and options. For Cygwin targets this is the default behavior.
16616 This option is deprecated and will be removed in a future release.
16619 @opindex mno-cygwin
16620 This option is available for Cygwin targets. It specifies that
16621 the MinGW internal interface is to be used instead of Cygwin's, by
16622 setting MinGW-related predefined macros and linker paths and default
16624 This option is deprecated and will be removed in a future release.
16628 This option is available for Cygwin and MinGW targets. It
16629 specifies that a DLL - a dynamic link library - is to be
16630 generated, enabling the selection of the required runtime
16631 startup object and entry point.
16633 @item -mnop-fun-dllimport
16634 @opindex mnop-fun-dllimport
16635 This option is available for Cygwin and MinGW targets. It
16636 specifies that the dllimport attribute should be ignored.
16640 This option is available for MinGW targets. It specifies
16641 that MinGW-specific thread support is to be used.
16645 This option is available for mingw-w64 targets. It specifies
16646 that the UNICODE macro is getting pre-defined and that the
16647 unicode capable runtime startup code is choosen.
16651 This option is available for Cygwin and MinGW targets. It
16652 specifies that the typical Windows pre-defined macros are to
16653 be set in the pre-processor, but does not influence the choice
16654 of runtime library/startup code.
16658 This option is available for Cygwin and MinGW targets. It
16659 specifies that a GUI application is to be generated by
16660 instructing the linker to set the PE header subsystem type
16663 @item -mpe-aligned-commons
16664 @opindex mpe-aligned-commons
16665 This option is available for Cygwin and MinGW targets. It
16666 specifies that the GNU extension to the PE file format that
16667 permits the correct alignment of COMMON variables should be
16668 used when generating code. It will be enabled by default if
16669 GCC detects that the target assembler found during configuration
16670 supports the feature.
16673 See also under @ref{i386 and x86-64 Options} for standard options.
16675 @node Xstormy16 Options
16676 @subsection Xstormy16 Options
16677 @cindex Xstormy16 Options
16679 These options are defined for Xstormy16:
16684 Choose startup files and linker script suitable for the simulator.
16687 @node Xtensa Options
16688 @subsection Xtensa Options
16689 @cindex Xtensa Options
16691 These options are supported for Xtensa targets:
16695 @itemx -mno-const16
16697 @opindex mno-const16
16698 Enable or disable use of @code{CONST16} instructions for loading
16699 constant values. The @code{CONST16} instruction is currently not a
16700 standard option from Tensilica. When enabled, @code{CONST16}
16701 instructions are always used in place of the standard @code{L32R}
16702 instructions. The use of @code{CONST16} is enabled by default only if
16703 the @code{L32R} instruction is not available.
16706 @itemx -mno-fused-madd
16707 @opindex mfused-madd
16708 @opindex mno-fused-madd
16709 Enable or disable use of fused multiply/add and multiply/subtract
16710 instructions in the floating-point option. This has no effect if the
16711 floating-point option is not also enabled. Disabling fused multiply/add
16712 and multiply/subtract instructions forces the compiler to use separate
16713 instructions for the multiply and add/subtract operations. This may be
16714 desirable in some cases where strict IEEE 754-compliant results are
16715 required: the fused multiply add/subtract instructions do not round the
16716 intermediate result, thereby producing results with @emph{more} bits of
16717 precision than specified by the IEEE standard. Disabling fused multiply
16718 add/subtract instructions also ensures that the program output is not
16719 sensitive to the compiler's ability to combine multiply and add/subtract
16722 @item -mserialize-volatile
16723 @itemx -mno-serialize-volatile
16724 @opindex mserialize-volatile
16725 @opindex mno-serialize-volatile
16726 When this option is enabled, GCC inserts @code{MEMW} instructions before
16727 @code{volatile} memory references to guarantee sequential consistency.
16728 The default is @option{-mserialize-volatile}. Use
16729 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
16731 @item -mtext-section-literals
16732 @itemx -mno-text-section-literals
16733 @opindex mtext-section-literals
16734 @opindex mno-text-section-literals
16735 Control the treatment of literal pools. The default is
16736 @option{-mno-text-section-literals}, which places literals in a separate
16737 section in the output file. This allows the literal pool to be placed
16738 in a data RAM/ROM, and it also allows the linker to combine literal
16739 pools from separate object files to remove redundant literals and
16740 improve code size. With @option{-mtext-section-literals}, the literals
16741 are interspersed in the text section in order to keep them as close as
16742 possible to their references. This may be necessary for large assembly
16745 @item -mtarget-align
16746 @itemx -mno-target-align
16747 @opindex mtarget-align
16748 @opindex mno-target-align
16749 When this option is enabled, GCC instructs the assembler to
16750 automatically align instructions to reduce branch penalties at the
16751 expense of some code density. The assembler attempts to widen density
16752 instructions to align branch targets and the instructions following call
16753 instructions. If there are not enough preceding safe density
16754 instructions to align a target, no widening will be performed. The
16755 default is @option{-mtarget-align}. These options do not affect the
16756 treatment of auto-aligned instructions like @code{LOOP}, which the
16757 assembler will always align, either by widening density instructions or
16758 by inserting no-op instructions.
16761 @itemx -mno-longcalls
16762 @opindex mlongcalls
16763 @opindex mno-longcalls
16764 When this option is enabled, GCC instructs the assembler to translate
16765 direct calls to indirect calls unless it can determine that the target
16766 of a direct call is in the range allowed by the call instruction. This
16767 translation typically occurs for calls to functions in other source
16768 files. Specifically, the assembler translates a direct @code{CALL}
16769 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
16770 The default is @option{-mno-longcalls}. This option should be used in
16771 programs where the call target can potentially be out of range. This
16772 option is implemented in the assembler, not the compiler, so the
16773 assembly code generated by GCC will still show direct call
16774 instructions---look at the disassembled object code to see the actual
16775 instructions. Note that the assembler will use an indirect call for
16776 every cross-file call, not just those that really will be out of range.
16779 @node zSeries Options
16780 @subsection zSeries Options
16781 @cindex zSeries options
16783 These are listed under @xref{S/390 and zSeries Options}.
16785 @node Code Gen Options
16786 @section Options for Code Generation Conventions
16787 @cindex code generation conventions
16788 @cindex options, code generation
16789 @cindex run-time options
16791 These machine-independent options control the interface conventions
16792 used in code generation.
16794 Most of them have both positive and negative forms; the negative form
16795 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
16796 one of the forms is listed---the one which is not the default. You
16797 can figure out the other form by either removing @samp{no-} or adding
16801 @item -fbounds-check
16802 @opindex fbounds-check
16803 For front-ends that support it, generate additional code to check that
16804 indices used to access arrays are within the declared range. This is
16805 currently only supported by the Java and Fortran front-ends, where
16806 this option defaults to true and false respectively.
16810 This option generates traps for signed overflow on addition, subtraction,
16811 multiplication operations.
16815 This option instructs the compiler to assume that signed arithmetic
16816 overflow of addition, subtraction and multiplication wraps around
16817 using twos-complement representation. This flag enables some optimizations
16818 and disables others. This option is enabled by default for the Java
16819 front-end, as required by the Java language specification.
16822 @opindex fexceptions
16823 Enable exception handling. Generates extra code needed to propagate
16824 exceptions. For some targets, this implies GCC will generate frame
16825 unwind information for all functions, which can produce significant data
16826 size overhead, although it does not affect execution. If you do not
16827 specify this option, GCC will enable it by default for languages like
16828 C++ which normally require exception handling, and disable it for
16829 languages like C that do not normally require it. However, you may need
16830 to enable this option when compiling C code that needs to interoperate
16831 properly with exception handlers written in C++. You may also wish to
16832 disable this option if you are compiling older C++ programs that don't
16833 use exception handling.
16835 @item -fnon-call-exceptions
16836 @opindex fnon-call-exceptions
16837 Generate code that allows trapping instructions to throw exceptions.
16838 Note that this requires platform-specific runtime support that does
16839 not exist everywhere. Moreover, it only allows @emph{trapping}
16840 instructions to throw exceptions, i.e.@: memory references or floating
16841 point instructions. It does not allow exceptions to be thrown from
16842 arbitrary signal handlers such as @code{SIGALRM}.
16844 @item -funwind-tables
16845 @opindex funwind-tables
16846 Similar to @option{-fexceptions}, except that it will just generate any needed
16847 static data, but will not affect the generated code in any other way.
16848 You will normally not enable this option; instead, a language processor
16849 that needs this handling would enable it on your behalf.
16851 @item -fasynchronous-unwind-tables
16852 @opindex fasynchronous-unwind-tables
16853 Generate unwind table in dwarf2 format, if supported by target machine. The
16854 table is exact at each instruction boundary, so it can be used for stack
16855 unwinding from asynchronous events (such as debugger or garbage collector).
16857 @item -fpcc-struct-return
16858 @opindex fpcc-struct-return
16859 Return ``short'' @code{struct} and @code{union} values in memory like
16860 longer ones, rather than in registers. This convention is less
16861 efficient, but it has the advantage of allowing intercallability between
16862 GCC-compiled files and files compiled with other compilers, particularly
16863 the Portable C Compiler (pcc).
16865 The precise convention for returning structures in memory depends
16866 on the target configuration macros.
16868 Short structures and unions are those whose size and alignment match
16869 that of some integer type.
16871 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
16872 switch is not binary compatible with code compiled with the
16873 @option{-freg-struct-return} switch.
16874 Use it to conform to a non-default application binary interface.
16876 @item -freg-struct-return
16877 @opindex freg-struct-return
16878 Return @code{struct} and @code{union} values in registers when possible.
16879 This is more efficient for small structures than
16880 @option{-fpcc-struct-return}.
16882 If you specify neither @option{-fpcc-struct-return} nor
16883 @option{-freg-struct-return}, GCC defaults to whichever convention is
16884 standard for the target. If there is no standard convention, GCC
16885 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
16886 the principal compiler. In those cases, we can choose the standard, and
16887 we chose the more efficient register return alternative.
16889 @strong{Warning:} code compiled with the @option{-freg-struct-return}
16890 switch is not binary compatible with code compiled with the
16891 @option{-fpcc-struct-return} switch.
16892 Use it to conform to a non-default application binary interface.
16894 @item -fshort-enums
16895 @opindex fshort-enums
16896 Allocate to an @code{enum} type only as many bytes as it needs for the
16897 declared range of possible values. Specifically, the @code{enum} type
16898 will be equivalent to the smallest integer type which has enough room.
16900 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
16901 code that is not binary compatible with code generated without that switch.
16902 Use it to conform to a non-default application binary interface.
16904 @item -fshort-double
16905 @opindex fshort-double
16906 Use the same size for @code{double} as for @code{float}.
16908 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
16909 code that is not binary compatible with code generated without that switch.
16910 Use it to conform to a non-default application binary interface.
16912 @item -fshort-wchar
16913 @opindex fshort-wchar
16914 Override the underlying type for @samp{wchar_t} to be @samp{short
16915 unsigned int} instead of the default for the target. This option is
16916 useful for building programs to run under WINE@.
16918 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
16919 code that is not binary compatible with code generated without that switch.
16920 Use it to conform to a non-default application binary interface.
16923 @opindex fno-common
16924 In C code, controls the placement of uninitialized global variables.
16925 Unix C compilers have traditionally permitted multiple definitions of
16926 such variables in different compilation units by placing the variables
16928 This is the behavior specified by @option{-fcommon}, and is the default
16929 for GCC on most targets.
16930 On the other hand, this behavior is not required by ISO C, and on some
16931 targets may carry a speed or code size penalty on variable references.
16932 The @option{-fno-common} option specifies that the compiler should place
16933 uninitialized global variables in the data section of the object file,
16934 rather than generating them as common blocks.
16935 This has the effect that if the same variable is declared
16936 (without @code{extern}) in two different compilations,
16937 you will get a multiple-definition error when you link them.
16938 In this case, you must compile with @option{-fcommon} instead.
16939 Compiling with @option{-fno-common} is useful on targets for which
16940 it provides better performance, or if you wish to verify that the
16941 program will work on other systems which always treat uninitialized
16942 variable declarations this way.
16946 Ignore the @samp{#ident} directive.
16948 @item -finhibit-size-directive
16949 @opindex finhibit-size-directive
16950 Don't output a @code{.size} assembler directive, or anything else that
16951 would cause trouble if the function is split in the middle, and the
16952 two halves are placed at locations far apart in memory. This option is
16953 used when compiling @file{crtstuff.c}; you should not need to use it
16956 @item -fverbose-asm
16957 @opindex fverbose-asm
16958 Put extra commentary information in the generated assembly code to
16959 make it more readable. This option is generally only of use to those
16960 who actually need to read the generated assembly code (perhaps while
16961 debugging the compiler itself).
16963 @option{-fno-verbose-asm}, the default, causes the
16964 extra information to be omitted and is useful when comparing two assembler
16967 @item -frecord-gcc-switches
16968 @opindex frecord-gcc-switches
16969 This switch causes the command line that was used to invoke the
16970 compiler to be recorded into the object file that is being created.
16971 This switch is only implemented on some targets and the exact format
16972 of the recording is target and binary file format dependent, but it
16973 usually takes the form of a section containing ASCII text. This
16974 switch is related to the @option{-fverbose-asm} switch, but that
16975 switch only records information in the assembler output file as
16976 comments, so it never reaches the object file.
16980 @cindex global offset table
16982 Generate position-independent code (PIC) suitable for use in a shared
16983 library, if supported for the target machine. Such code accesses all
16984 constant addresses through a global offset table (GOT)@. The dynamic
16985 loader resolves the GOT entries when the program starts (the dynamic
16986 loader is not part of GCC; it is part of the operating system). If
16987 the GOT size for the linked executable exceeds a machine-specific
16988 maximum size, you get an error message from the linker indicating that
16989 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
16990 instead. (These maximums are 8k on the SPARC and 32k
16991 on the m68k and RS/6000. The 386 has no such limit.)
16993 Position-independent code requires special support, and therefore works
16994 only on certain machines. For the 386, GCC supports PIC for System V
16995 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
16996 position-independent.
16998 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17003 If supported for the target machine, emit position-independent code,
17004 suitable for dynamic linking and avoiding any limit on the size of the
17005 global offset table. This option makes a difference on the m68k,
17006 PowerPC and SPARC@.
17008 Position-independent code requires special support, and therefore works
17009 only on certain machines.
17011 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17018 These options are similar to @option{-fpic} and @option{-fPIC}, but
17019 generated position independent code can be only linked into executables.
17020 Usually these options are used when @option{-pie} GCC option will be
17021 used during linking.
17023 @option{-fpie} and @option{-fPIE} both define the macros
17024 @code{__pie__} and @code{__PIE__}. The macros have the value 1
17025 for @option{-fpie} and 2 for @option{-fPIE}.
17027 @item -fno-jump-tables
17028 @opindex fno-jump-tables
17029 Do not use jump tables for switch statements even where it would be
17030 more efficient than other code generation strategies. This option is
17031 of use in conjunction with @option{-fpic} or @option{-fPIC} for
17032 building code which forms part of a dynamic linker and cannot
17033 reference the address of a jump table. On some targets, jump tables
17034 do not require a GOT and this option is not needed.
17036 @item -ffixed-@var{reg}
17038 Treat the register named @var{reg} as a fixed register; generated code
17039 should never refer to it (except perhaps as a stack pointer, frame
17040 pointer or in some other fixed role).
17042 @var{reg} must be the name of a register. The register names accepted
17043 are machine-specific and are defined in the @code{REGISTER_NAMES}
17044 macro in the machine description macro file.
17046 This flag does not have a negative form, because it specifies a
17049 @item -fcall-used-@var{reg}
17050 @opindex fcall-used
17051 Treat the register named @var{reg} as an allocable register that is
17052 clobbered by function calls. It may be allocated for temporaries or
17053 variables that do not live across a call. Functions compiled this way
17054 will not save and restore the register @var{reg}.
17056 It is an error to used this flag with the frame pointer or stack pointer.
17057 Use of this flag for other registers that have fixed pervasive roles in
17058 the machine's execution model will produce disastrous results.
17060 This flag does not have a negative form, because it specifies a
17063 @item -fcall-saved-@var{reg}
17064 @opindex fcall-saved
17065 Treat the register named @var{reg} as an allocable register saved by
17066 functions. It may be allocated even for temporaries or variables that
17067 live across a call. Functions compiled this way will save and restore
17068 the register @var{reg} if they use it.
17070 It is an error to used this flag with the frame pointer or stack pointer.
17071 Use of this flag for other registers that have fixed pervasive roles in
17072 the machine's execution model will produce disastrous results.
17074 A different sort of disaster will result from the use of this flag for
17075 a register in which function values may be returned.
17077 This flag does not have a negative form, because it specifies a
17080 @item -fpack-struct[=@var{n}]
17081 @opindex fpack-struct
17082 Without a value specified, pack all structure members together without
17083 holes. When a value is specified (which must be a small power of two), pack
17084 structure members according to this value, representing the maximum
17085 alignment (that is, objects with default alignment requirements larger than
17086 this will be output potentially unaligned at the next fitting location.
17088 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
17089 code that is not binary compatible with code generated without that switch.
17090 Additionally, it makes the code suboptimal.
17091 Use it to conform to a non-default application binary interface.
17093 @item -finstrument-functions
17094 @opindex finstrument-functions
17095 Generate instrumentation calls for entry and exit to functions. Just
17096 after function entry and just before function exit, the following
17097 profiling functions will be called with the address of the current
17098 function and its call site. (On some platforms,
17099 @code{__builtin_return_address} does not work beyond the current
17100 function, so the call site information may not be available to the
17101 profiling functions otherwise.)
17104 void __cyg_profile_func_enter (void *this_fn,
17106 void __cyg_profile_func_exit (void *this_fn,
17110 The first argument is the address of the start of the current function,
17111 which may be looked up exactly in the symbol table.
17113 This instrumentation is also done for functions expanded inline in other
17114 functions. The profiling calls will indicate where, conceptually, the
17115 inline function is entered and exited. This means that addressable
17116 versions of such functions must be available. If all your uses of a
17117 function are expanded inline, this may mean an additional expansion of
17118 code size. If you use @samp{extern inline} in your C code, an
17119 addressable version of such functions must be provided. (This is
17120 normally the case anyways, but if you get lucky and the optimizer always
17121 expands the functions inline, you might have gotten away without
17122 providing static copies.)
17124 A function may be given the attribute @code{no_instrument_function}, in
17125 which case this instrumentation will not be done. This can be used, for
17126 example, for the profiling functions listed above, high-priority
17127 interrupt routines, and any functions from which the profiling functions
17128 cannot safely be called (perhaps signal handlers, if the profiling
17129 routines generate output or allocate memory).
17131 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
17132 @opindex finstrument-functions-exclude-file-list
17134 Set the list of functions that are excluded from instrumentation (see
17135 the description of @code{-finstrument-functions}). If the file that
17136 contains a function definition matches with one of @var{file}, then
17137 that function is not instrumented. The match is done on substrings:
17138 if the @var{file} parameter is a substring of the file name, it is
17139 considered to be a match.
17142 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
17143 will exclude any inline function defined in files whose pathnames
17144 contain @code{/bits/stl} or @code{include/sys}.
17146 If, for some reason, you want to include letter @code{','} in one of
17147 @var{sym}, write @code{'\,'}. For example,
17148 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
17149 (note the single quote surrounding the option).
17151 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
17152 @opindex finstrument-functions-exclude-function-list
17154 This is similar to @code{-finstrument-functions-exclude-file-list},
17155 but this option sets the list of function names to be excluded from
17156 instrumentation. The function name to be matched is its user-visible
17157 name, such as @code{vector<int> blah(const vector<int> &)}, not the
17158 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
17159 match is done on substrings: if the @var{sym} parameter is a substring
17160 of the function name, it is considered to be a match. For C99 and C++
17161 extended identifiers, the function name must be given in UTF-8, not
17162 using universal character names.
17164 @item -fstack-check
17165 @opindex fstack-check
17166 Generate code to verify that you do not go beyond the boundary of the
17167 stack. You should specify this flag if you are running in an
17168 environment with multiple threads, but only rarely need to specify it in
17169 a single-threaded environment since stack overflow is automatically
17170 detected on nearly all systems if there is only one stack.
17172 Note that this switch does not actually cause checking to be done; the
17173 operating system or the language runtime must do that. The switch causes
17174 generation of code to ensure that they see the stack being extended.
17176 You can additionally specify a string parameter: @code{no} means no
17177 checking, @code{generic} means force the use of old-style checking,
17178 @code{specific} means use the best checking method and is equivalent
17179 to bare @option{-fstack-check}.
17181 Old-style checking is a generic mechanism that requires no specific
17182 target support in the compiler but comes with the following drawbacks:
17186 Modified allocation strategy for large objects: they will always be
17187 allocated dynamically if their size exceeds a fixed threshold.
17190 Fixed limit on the size of the static frame of functions: when it is
17191 topped by a particular function, stack checking is not reliable and
17192 a warning is issued by the compiler.
17195 Inefficiency: because of both the modified allocation strategy and the
17196 generic implementation, the performances of the code are hampered.
17199 Note that old-style stack checking is also the fallback method for
17200 @code{specific} if no target support has been added in the compiler.
17202 @item -fstack-limit-register=@var{reg}
17203 @itemx -fstack-limit-symbol=@var{sym}
17204 @itemx -fno-stack-limit
17205 @opindex fstack-limit-register
17206 @opindex fstack-limit-symbol
17207 @opindex fno-stack-limit
17208 Generate code to ensure that the stack does not grow beyond a certain value,
17209 either the value of a register or the address of a symbol. If the stack
17210 would grow beyond the value, a signal is raised. For most targets,
17211 the signal is raised before the stack overruns the boundary, so
17212 it is possible to catch the signal without taking special precautions.
17214 For instance, if the stack starts at absolute address @samp{0x80000000}
17215 and grows downwards, you can use the flags
17216 @option{-fstack-limit-symbol=__stack_limit} and
17217 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
17218 of 128KB@. Note that this may only work with the GNU linker.
17220 @cindex aliasing of parameters
17221 @cindex parameters, aliased
17222 @item -fargument-alias
17223 @itemx -fargument-noalias
17224 @itemx -fargument-noalias-global
17225 @itemx -fargument-noalias-anything
17226 @opindex fargument-alias
17227 @opindex fargument-noalias
17228 @opindex fargument-noalias-global
17229 @opindex fargument-noalias-anything
17230 Specify the possible relationships among parameters and between
17231 parameters and global data.
17233 @option{-fargument-alias} specifies that arguments (parameters) may
17234 alias each other and may alias global storage.@*
17235 @option{-fargument-noalias} specifies that arguments do not alias
17236 each other, but may alias global storage.@*
17237 @option{-fargument-noalias-global} specifies that arguments do not
17238 alias each other and do not alias global storage.
17239 @option{-fargument-noalias-anything} specifies that arguments do not
17240 alias any other storage.
17242 Each language will automatically use whatever option is required by
17243 the language standard. You should not need to use these options yourself.
17245 @item -fleading-underscore
17246 @opindex fleading-underscore
17247 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
17248 change the way C symbols are represented in the object file. One use
17249 is to help link with legacy assembly code.
17251 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
17252 generate code that is not binary compatible with code generated without that
17253 switch. Use it to conform to a non-default application binary interface.
17254 Not all targets provide complete support for this switch.
17256 @item -ftls-model=@var{model}
17257 @opindex ftls-model
17258 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
17259 The @var{model} argument should be one of @code{global-dynamic},
17260 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
17262 The default without @option{-fpic} is @code{initial-exec}; with
17263 @option{-fpic} the default is @code{global-dynamic}.
17265 @item -fvisibility=@var{default|internal|hidden|protected}
17266 @opindex fvisibility
17267 Set the default ELF image symbol visibility to the specified option---all
17268 symbols will be marked with this unless overridden within the code.
17269 Using this feature can very substantially improve linking and
17270 load times of shared object libraries, produce more optimized
17271 code, provide near-perfect API export and prevent symbol clashes.
17272 It is @strong{strongly} recommended that you use this in any shared objects
17275 Despite the nomenclature, @code{default} always means public ie;
17276 available to be linked against from outside the shared object.
17277 @code{protected} and @code{internal} are pretty useless in real-world
17278 usage so the only other commonly used option will be @code{hidden}.
17279 The default if @option{-fvisibility} isn't specified is
17280 @code{default}, i.e., make every
17281 symbol public---this causes the same behavior as previous versions of
17284 A good explanation of the benefits offered by ensuring ELF
17285 symbols have the correct visibility is given by ``How To Write
17286 Shared Libraries'' by Ulrich Drepper (which can be found at
17287 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
17288 solution made possible by this option to marking things hidden when
17289 the default is public is to make the default hidden and mark things
17290 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
17291 and @code{__attribute__ ((visibility("default")))} instead of
17292 @code{__declspec(dllexport)} you get almost identical semantics with
17293 identical syntax. This is a great boon to those working with
17294 cross-platform projects.
17296 For those adding visibility support to existing code, you may find
17297 @samp{#pragma GCC visibility} of use. This works by you enclosing
17298 the declarations you wish to set visibility for with (for example)
17299 @samp{#pragma GCC visibility push(hidden)} and
17300 @samp{#pragma GCC visibility pop}.
17301 Bear in mind that symbol visibility should be viewed @strong{as
17302 part of the API interface contract} and thus all new code should
17303 always specify visibility when it is not the default ie; declarations
17304 only for use within the local DSO should @strong{always} be marked explicitly
17305 as hidden as so to avoid PLT indirection overheads---making this
17306 abundantly clear also aids readability and self-documentation of the code.
17307 Note that due to ISO C++ specification requirements, operator new and
17308 operator delete must always be of default visibility.
17310 Be aware that headers from outside your project, in particular system
17311 headers and headers from any other library you use, may not be
17312 expecting to be compiled with visibility other than the default. You
17313 may need to explicitly say @samp{#pragma GCC visibility push(default)}
17314 before including any such headers.
17316 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
17317 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
17318 no modifications. However, this means that calls to @samp{extern}
17319 functions with no explicit visibility will use the PLT, so it is more
17320 effective to use @samp{__attribute ((visibility))} and/or
17321 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
17322 declarations should be treated as hidden.
17324 Note that @samp{-fvisibility} does affect C++ vague linkage
17325 entities. This means that, for instance, an exception class that will
17326 be thrown between DSOs must be explicitly marked with default
17327 visibility so that the @samp{type_info} nodes will be unified between
17330 An overview of these techniques, their benefits and how to use them
17331 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
17337 @node Environment Variables
17338 @section Environment Variables Affecting GCC
17339 @cindex environment variables
17341 @c man begin ENVIRONMENT
17342 This section describes several environment variables that affect how GCC
17343 operates. Some of them work by specifying directories or prefixes to use
17344 when searching for various kinds of files. Some are used to specify other
17345 aspects of the compilation environment.
17347 Note that you can also specify places to search using options such as
17348 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
17349 take precedence over places specified using environment variables, which
17350 in turn take precedence over those specified by the configuration of GCC@.
17351 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
17352 GNU Compiler Collection (GCC) Internals}.
17357 @c @itemx LC_COLLATE
17359 @c @itemx LC_MONETARY
17360 @c @itemx LC_NUMERIC
17365 @c @findex LC_COLLATE
17366 @findex LC_MESSAGES
17367 @c @findex LC_MONETARY
17368 @c @findex LC_NUMERIC
17372 These environment variables control the way that GCC uses
17373 localization information that allow GCC to work with different
17374 national conventions. GCC inspects the locale categories
17375 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
17376 so. These locale categories can be set to any value supported by your
17377 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
17378 Kingdom encoded in UTF-8.
17380 The @env{LC_CTYPE} environment variable specifies character
17381 classification. GCC uses it to determine the character boundaries in
17382 a string; this is needed for some multibyte encodings that contain quote
17383 and escape characters that would otherwise be interpreted as a string
17386 The @env{LC_MESSAGES} environment variable specifies the language to
17387 use in diagnostic messages.
17389 If the @env{LC_ALL} environment variable is set, it overrides the value
17390 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
17391 and @env{LC_MESSAGES} default to the value of the @env{LANG}
17392 environment variable. If none of these variables are set, GCC
17393 defaults to traditional C English behavior.
17397 If @env{TMPDIR} is set, it specifies the directory to use for temporary
17398 files. GCC uses temporary files to hold the output of one stage of
17399 compilation which is to be used as input to the next stage: for example,
17400 the output of the preprocessor, which is the input to the compiler
17403 @item GCC_EXEC_PREFIX
17404 @findex GCC_EXEC_PREFIX
17405 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
17406 names of the subprograms executed by the compiler. No slash is added
17407 when this prefix is combined with the name of a subprogram, but you can
17408 specify a prefix that ends with a slash if you wish.
17410 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
17411 an appropriate prefix to use based on the pathname it was invoked with.
17413 If GCC cannot find the subprogram using the specified prefix, it
17414 tries looking in the usual places for the subprogram.
17416 The default value of @env{GCC_EXEC_PREFIX} is
17417 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
17418 the installed compiler. In many cases @var{prefix} is the value
17419 of @code{prefix} when you ran the @file{configure} script.
17421 Other prefixes specified with @option{-B} take precedence over this prefix.
17423 This prefix is also used for finding files such as @file{crt0.o} that are
17426 In addition, the prefix is used in an unusual way in finding the
17427 directories to search for header files. For each of the standard
17428 directories whose name normally begins with @samp{/usr/local/lib/gcc}
17429 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
17430 replacing that beginning with the specified prefix to produce an
17431 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
17432 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
17433 These alternate directories are searched first; the standard directories
17434 come next. If a standard directory begins with the configured
17435 @var{prefix} then the value of @var{prefix} is replaced by
17436 @env{GCC_EXEC_PREFIX} when looking for header files.
17438 @item COMPILER_PATH
17439 @findex COMPILER_PATH
17440 The value of @env{COMPILER_PATH} is a colon-separated list of
17441 directories, much like @env{PATH}. GCC tries the directories thus
17442 specified when searching for subprograms, if it can't find the
17443 subprograms using @env{GCC_EXEC_PREFIX}.
17446 @findex LIBRARY_PATH
17447 The value of @env{LIBRARY_PATH} is a colon-separated list of
17448 directories, much like @env{PATH}. When configured as a native compiler,
17449 GCC tries the directories thus specified when searching for special
17450 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
17451 using GCC also uses these directories when searching for ordinary
17452 libraries for the @option{-l} option (but directories specified with
17453 @option{-L} come first).
17457 @cindex locale definition
17458 This variable is used to pass locale information to the compiler. One way in
17459 which this information is used is to determine the character set to be used
17460 when character literals, string literals and comments are parsed in C and C++.
17461 When the compiler is configured to allow multibyte characters,
17462 the following values for @env{LANG} are recognized:
17466 Recognize JIS characters.
17468 Recognize SJIS characters.
17470 Recognize EUCJP characters.
17473 If @env{LANG} is not defined, or if it has some other value, then the
17474 compiler will use mblen and mbtowc as defined by the default locale to
17475 recognize and translate multibyte characters.
17479 Some additional environments variables affect the behavior of the
17482 @include cppenv.texi
17486 @node Precompiled Headers
17487 @section Using Precompiled Headers
17488 @cindex precompiled headers
17489 @cindex speed of compilation
17491 Often large projects have many header files that are included in every
17492 source file. The time the compiler takes to process these header files
17493 over and over again can account for nearly all of the time required to
17494 build the project. To make builds faster, GCC allows users to
17495 `precompile' a header file; then, if builds can use the precompiled
17496 header file they will be much faster.
17498 To create a precompiled header file, simply compile it as you would any
17499 other file, if necessary using the @option{-x} option to make the driver
17500 treat it as a C or C++ header file. You will probably want to use a
17501 tool like @command{make} to keep the precompiled header up-to-date when
17502 the headers it contains change.
17504 A precompiled header file will be searched for when @code{#include} is
17505 seen in the compilation. As it searches for the included file
17506 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
17507 compiler looks for a precompiled header in each directory just before it
17508 looks for the include file in that directory. The name searched for is
17509 the name specified in the @code{#include} with @samp{.gch} appended. If
17510 the precompiled header file can't be used, it is ignored.
17512 For instance, if you have @code{#include "all.h"}, and you have
17513 @file{all.h.gch} in the same directory as @file{all.h}, then the
17514 precompiled header file will be used if possible, and the original
17515 header will be used otherwise.
17517 Alternatively, you might decide to put the precompiled header file in a
17518 directory and use @option{-I} to ensure that directory is searched
17519 before (or instead of) the directory containing the original header.
17520 Then, if you want to check that the precompiled header file is always
17521 used, you can put a file of the same name as the original header in this
17522 directory containing an @code{#error} command.
17524 This also works with @option{-include}. So yet another way to use
17525 precompiled headers, good for projects not designed with precompiled
17526 header files in mind, is to simply take most of the header files used by
17527 a project, include them from another header file, precompile that header
17528 file, and @option{-include} the precompiled header. If the header files
17529 have guards against multiple inclusion, they will be skipped because
17530 they've already been included (in the precompiled header).
17532 If you need to precompile the same header file for different
17533 languages, targets, or compiler options, you can instead make a
17534 @emph{directory} named like @file{all.h.gch}, and put each precompiled
17535 header in the directory, perhaps using @option{-o}. It doesn't matter
17536 what you call the files in the directory, every precompiled header in
17537 the directory will be considered. The first precompiled header
17538 encountered in the directory that is valid for this compilation will
17539 be used; they're searched in no particular order.
17541 There are many other possibilities, limited only by your imagination,
17542 good sense, and the constraints of your build system.
17544 A precompiled header file can be used only when these conditions apply:
17548 Only one precompiled header can be used in a particular compilation.
17551 A precompiled header can't be used once the first C token is seen. You
17552 can have preprocessor directives before a precompiled header; you can
17553 even include a precompiled header from inside another header, so long as
17554 there are no C tokens before the @code{#include}.
17557 The precompiled header file must be produced for the same language as
17558 the current compilation. You can't use a C precompiled header for a C++
17562 The precompiled header file must have been produced by the same compiler
17563 binary as the current compilation is using.
17566 Any macros defined before the precompiled header is included must
17567 either be defined in the same way as when the precompiled header was
17568 generated, or must not affect the precompiled header, which usually
17569 means that they don't appear in the precompiled header at all.
17571 The @option{-D} option is one way to define a macro before a
17572 precompiled header is included; using a @code{#define} can also do it.
17573 There are also some options that define macros implicitly, like
17574 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
17577 @item If debugging information is output when using the precompiled
17578 header, using @option{-g} or similar, the same kind of debugging information
17579 must have been output when building the precompiled header. However,
17580 a precompiled header built using @option{-g} can be used in a compilation
17581 when no debugging information is being output.
17583 @item The same @option{-m} options must generally be used when building
17584 and using the precompiled header. @xref{Submodel Options},
17585 for any cases where this rule is relaxed.
17587 @item Each of the following options must be the same when building and using
17588 the precompiled header:
17590 @gccoptlist{-fexceptions}
17593 Some other command-line options starting with @option{-f},
17594 @option{-p}, or @option{-O} must be defined in the same way as when
17595 the precompiled header was generated. At present, it's not clear
17596 which options are safe to change and which are not; the safest choice
17597 is to use exactly the same options when generating and using the
17598 precompiled header. The following are known to be safe:
17600 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
17601 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
17602 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
17607 For all of these except the last, the compiler will automatically
17608 ignore the precompiled header if the conditions aren't met. If you
17609 find an option combination that doesn't work and doesn't cause the
17610 precompiled header to be ignored, please consider filing a bug report,
17613 If you do use differing options when generating and using the
17614 precompiled header, the actual behavior will be a mixture of the
17615 behavior for the options. For instance, if you use @option{-g} to
17616 generate the precompiled header but not when using it, you may or may
17617 not get debugging information for routines in the precompiled header.